Gentoo elections are conducted using a custom software called votify. During the voting period, the developers place their votes in their respective home directories on one of the Gentoo servers. Afterwards, the election officials collect the votes, count them, compare their results and finally announce them.

The simplified description stated above suggests two weak points. Firstly, we rely on honesty of election officials. If they chose to conspire, they could fake the result. Secondly, we rely on honesty of all Infrastructure members, as they could use root access to manipulate the votes (or the collection process).

To protect against possible fraud, we make the elections transparent (but pseudonymous). This means that all votes cast are public, so everyone can count them and verify the result. Furthermore, developers can verify whether their personal vote has been included. Ideally, all developers would do that and therefore confirm that no votes were manipulated.

Currently, we are pretty much implicitly relying on developers doing that, and assuming that no protest implies successful verification. However, this is not really reliable, and given the unfriendly nature of our scripts I have reasons to doubt that the majority of developers actually verify the election results. In this post, I would like to shortly explain how Gentoo elections work, how they could be manipulated and introduce Votrify — a tool to explicitly verify election results.

Gentoo voting process in detail

Once the nomination period is over, an election official sets the voting process up by creating control files for the voting scripts. Those control files include election name, voting period, ballot (containing all vote choices) and list of eligible voters.

There are no explicit events corresponding to the beginning or the end of voting period. The votify script used by developers reads election data on each execution, and uses it to determine whether the voting period is open. During the voting period, it permits the developer to edit the vote, and finally to ‘submit’ it. Both draft and submitted vote are stored as appropriate files in the developer’s home directory, ‘submitted’ votes are not collected automatically. This means that the developer can still manually manipulate the vote once voting period concludes, and before the votes are manually collected.

Votes are collected explicitly by an election official. When run, the countify script collects all vote files from developers’ home directories. An unique ‘confirmation ID’ is generated for each voting developer. All votes along with their confirmation IDs are placed in so-called ‘master ballot’, while mapping from developer names to confirmation IDs is stored separately. The latter is used to send developers their respective confirmation IDs, and can be discarded afterwards.

Each of the election officials uses the master ballot to count the votes. Afterwards, they compare their results and if they match, they announce the election results. The master ballot is attached to the announcement mail, so that everyone can verify the results.

Possible manipulations

The three methods of manipulating the vote that I can think of are:

1. Announcing fake results. An election result may be presented that does not match the votes cast. This is actively prevented by having multiple election officials, and by making the votes transparent so that everyone can count them.
2. Manipulating votes cast by developers. The result could be manipulated by modifying the votes cast by individual developers. This is prevented by including pseudonymous vote attribution in the master ballot. Every developer can therefore check whether his/her vote has been reproduced correctly. However, this presumes that the developer is active.
3. Adding fake votes to the master ballot. The result could be manipulated by adding votes that were not cast by any of the existing developers. This is a major problem, and such manipulation is entirely plausible if the turnout is low enough, and developers who did not vote fail to check whether they have not been added to the casting voter list.

Furthermore, the efficiency of the last method can be improved if the attacker is able to restrict communication between voters and/or reliably deliver different versions of the master ballot to different voters, i.e. convince the voters that their own vote was included correctly while manipulating the remaining votes to achieve the desired result. The former is rather unlikely but the latter is generally feasible.

Finally, the results could be manipulated via manipulating the voting software. This can be counteracted through verifying the implementation against the algorithm specification or, to some degree, via comparing the results a third party tool. Robin H. Johnson and myself were historically working on this (or more specifically, on verifying whether the Gentoo implementation of Schulze method is correct) but neither of us was able to finish the work. If you’re interested in the topic, you can look at my election-compare repository. For the purpose of this post, I’m going to consider this possibility out of scope.

Verifying election results using Votrify

Votrify uses a two-stage verification model. It consists of individual verification which is performed by each voter separately and produces signed confirmations, and community verification that uses the aforementioned files to provide final verified election result.

The individual verification part involves:

1. Verifying that the developer’s vote has been recorded correctly. This takes part in detecting whether any votes have been manipulated. The positive result of this verification is implied by the fact that a confirmation is produced. Additionally, developers who did not cast a vote also need to produce confirmations, in order to detect any extraneous votes.
2. Counting the votes and producing the election result. This produces the election results as seen from the developer’s perspective, and therefore prevents manipulation via announcing fake results. Furthermore, comparing the results between different developers helps finding implementation bugs.
3. Hashing the master ballot. The hash of master ballot file is included, and comparing it between different results confirms that all voters received the same master ballot.

If the verification is positive, a confirmation is produced and signed using developer’s OpenPGP key. I would like to note that no private data is leaked in the process. It does not even indicate whether the dev in question has actually voted — only that he/she participates in the verification process.

Afterwards, confirmations from different voters are collected. They are used to perform community verification which involves:

1. Verifying the OpenPGP signature. This is necessary to confirm the authenticity of the signed confirmation. The check also involves verifying that the key owner was an eligible voter and that each voter produced only one confirmation. Therefore, it prevents attempts to~fake the verification results.
2. Comparing the results and master ballot hashes. This confirms that everyone participating received the same master ballot, and produced the same results.

If the verification for all confirmations is positive, the election results are repeated, along with explicit quantification of how trustworthy they are. The number indicates how many confirmations were used, and therefore how many of the votes (or non-votes) in master ballot were confirmed. The difference between the number of eligible voters and the number of confirmations indicates how many votes may have been altered, planted or deleted. Ideally, if all eligible voters produced signed confirmations, the election would be 100% confirmed.

The recent key poisoning attack on SKS keyservers shook the world of OpenPGP. While this isn’t a new problem, it has not been exploited on this scale before. The attackers have proved how easy it is to poison commonly used keys on the keyservers and effectively render GnuPG unusably slow. A renewed discussion on improving keyservers has started as a result. It also forced Gentoo to employ countermeasures. You can read more on them in the ‘Impact of SKS keyserver poisoning on Gentoo’ news item.

Coincidentally, the attack happened shortly after the launch of keys.openpgp.org, that advertises itself as both poisoning-resistant and GDPR-friendly keyserver. Naturally, many users see it as the ultimate solution to the issues with SKS. I’m afraid I have to disagree — in my opinion, this keyserver does not solve any problems, it merely cripples OpenPGP in order to avoid being affected by them, and harms its security in the process.

In this article, I’d like to shortly explain what the problem is, and which of the different solutions proposed so far to it (e.g. on gnupg-users mailing list) make sense, and which make things even worse. Naturally, I will also cover the new Hagrid keyserver as one of the glorified non-solutions.

The attack — key poisoning

OpenPGP uses a distributed design — once the primary key is created, additional packets can be freely appended to it and recombined on different systems. Those packets include subkeys, user identifiers and signatures. Signatures are used to confirm the authenticity of appended packets. The packets are only meaningful if the client can verify the authenticity of their respective signatures.

The attack is carried through third-party signatures that normally are used by different people to confirm the authenticity of the key — that is, to state that the signer has verified the identity of the key owner. It relies on three distinct properties of OpenPGP:

1. The key can contain unlimited number of signatures. After all, it is natural that very old keys will have a large number of signatures made by different people on them.
2. Anyone can append signatures to any OpenPGP key. This is partially keyserver policy, and partially the fact that SKS keyserver nodes are propagating keys one to another.
3. There is no way to distinguish legitimate signatures from garbage. To put it other way, it is trivial to make garbage signatures look like the real deal.

The attacker abuses those properties by creating a large number of garbage signatures and sending them to keyservers. When users fetch key updates from the keyserver, GnuPG normally appends all those signatures to the local copy. As a result, the key becomes unusually large and causes severe performance issues with GnuPG, preventing its normal usage. The user ends up having to manually remove the key in order to fix the installation.

The obvious non-solutions and potential solutions

Let’s start by analyzing the properties I’ve listed above. After all, removing at least one of the requirements should prevent the attack from being possible. But can we really do that?

Firstly, we could set a hard limit on number of signatures or key size. This should obviously prevent the attacker from breaking user systems via huge keys. However, it will make it entirely possible for the attacker to ‘brick’ the key by appending garbage up to the limit. Then it would no longer be possible to append any valid signatures to the key. Users would suffer less but the key owner will lose the ability to use the key meaningfully. It’s a no-go.

Secondly, we could limit key updates to the owner. However, the keyserver update protocol currently does not provide any standard way of verifying who the uploader is, so it would effectively require incompatible changes at least to the upload protocol. Furthermore, in order to prevent malicious keyservers from propagating fake signatures we’d also need to carry the verification along when propagating key updates. This effectively means an extension of the key format, and it has been proposed e.g. in ‘Abuse-Resistant OpenPGP Keystores’ draft. This is probably a wortwhile option but it will take time before it’s implemented.

Thirdly, we could try to validate signatures. However, any validation can be easily worked around. If we started requiring signing keys to be present on the keyserver, the attackers can simply mass-upload keys used to create garbage signatures. If we went even further and e.g. started requiring verified e-mail addresses for the signing keys, the attackers can simply mass-create e-mail addresses and verify them. It might work as a temporary solution but it will probably cause more harm than good.

There were other non-solutions suggested — most notably, blacklisting poisoned keys. However, this is even worse. It means that every victim of poisoning attack would be excluded from using the keyserver, and in my opinion it will only provoke the attackers to poison even more keys. It may sound like a good interim solution preventing users from being hit but it is rather short-sighted.

keys.openpgp.org / Hagrid — a big non-solution

A common suggestion for OpenPGP users — one that even Gentoo news item mentions for lack of alternative — is to switch to keys.openpgp.org keyserver, or switch keyservers to their Hagrid software. It is not vulnerable to key poisoning attack because it strips away all third-party signatures. However, this and other limitations make it a rather poor replacement, and in my opinion can be harmful to security of OpenPGP.

Firstly, stripping all third-party signatures is not a solution. It simply avoids the problem by killing a very important portion of OpenPGP protocol — the Web of Trust. Without it, the keys obtained from the server can not be authenticated otherwise than by direct interaction between the individuals. For example, Gentoo Authority Keys can’t work there. Most of the time, you won’t be able to tell whether the key on keyserver is legitimate or forged.

The e-mail verification makes it even worse, though not intentionally. While I agree that many users do not understand or use WoT, Hagrid is implicitly going to cause users to start relying on e-mail verification as proof of key authenticity. In other words, people are going to assume that if a key on keys.openpgp.org has verified e-mail address, it has to be legitimate. This makes it trivial for an attacker that manages to gain unauthorized access to the e-mail address or the keyserver to publish a forged key and convince others to use it.

Secondly, Hagrid does not support UID revocations. This is an entirely absurd case where GDPR fear won over security. If your e-mail address becomes compromised, you will not be able to revoke it. Sure, the keyserver admins may eventually stop propagating it along with your key, but all users who fetched the key before will continue seeing it as a valid UID. Of course, if users send encrypted mail the attacker won’t be able to read it. However, the users can be trivially persuaded to switch to a new, forged key.

Thirdly, Hagrid rejects all UIDs except for verified e-mail-based UIDs. This is something we could live with if key owners actively pursue having their identities verified. However, this also means you can’t publish a photo identity or use keybase.io. The ‘explicit consent’ argument used by upstream is rather silly — apparently every UID requires separate consent, while at the same time you can trivially share somebody else’s PII as the real name of a valid e-mail address.

Apparently, upstream is willing to resolve the first two of those issues once satisfactory solutions are established. However, this doesn’t mean that it’s fine to ignore those problems. Until they are resolved, and necessary OpenPGP client updates are sufficiently widely deployed, I don’t believe Hagrid or its instance at keys.openpgp.org are good replacements for SKS and other keyservers.

So what are the solutions?

Sadly, I am not aware of any good global solution at the moment. The best workaround for GnuPG users so far is the new self-sigs-only option that prevents it from importing third-party signatures. Of course, it shares the first limitation of Hagrid keyserver. The future versions of GnuPG will supposedly fallback to this option upon meeting excessively large keys.

For domain-limited use cases such as Gentoo’s, running a local keyserver with restricted upload access is an option. However, it requires users to explicitly specify our keyserver, and effectively end up having to specify multiple different keyservers for each domain. Furthermore, WKD can be used to distribute keys. Sadly, at the moment GnuPG uses it only to locate new keys and does not support refreshing keys via WKD (gemato employs a cheap hack to make it happen). In both cases, the attack is prevented via isolating the infrastructure and preventing public upload access.

The long-term solution probably lies in the ‘First-party-attested Third-party Certifications‘ section of the ‘Abuse-Resistant OpenPGP Keystores’ draft. In this proposal, every third-party signature must be explicitly attested by the key owner. Therefore, only the key owner can append additional signatures to the key, and keyservers can reject any signatures that were not attested. However, this is not currently supported by GnuPG, and once it is, deploying it will most likely take significant time.

tl;dr: Here (PDF, 67 kB) is a case label for Pocket Science Lab version 5 that is compatible with the design for a laser-cut case published by FOSSAsia.

In case you haven’t heard about it, Pocket Science Lab [1] is a really nifty board developed by the FOSSAsia community which combines a multichannel, megahertz-range oscilloscope, a multimeter, a logic probe, several voltage sources and a current source, several wave generators, UART and I2C interfaces… and all of this in the form factor of an Arduino Mega, i.e. only somewhat larger than that of a credit card. Hook it up over USB to a PC or an Android device running the official (free and open source, of course) app and you are all set.

Well, not quite set yet. What you get for your 50-ish EUR is just the board itself. You will quite definitely need a set of probe cables (sadly, I have yet to find even an unofficial adaptor allowing one to equip PSLab with standard industry oscilloscope probes using BNC connectors) but if you expect to lug yours around anywhere you go, you will quite definitely want to invest in a case of some sort. While FOSSAsia does not to my knowledge sell PSLab cases, they provide a design for one [2]. It is meant to be laser-cut but I have successfully managed to 3D-print it as well, and for the more patient among us it shouldn’t be too difficult to hand-cut one with a jigsaw either.

Of course in addition to making sure your Pocket Science Lab is protected against accidental damage it would also be nice to have all the connectors clearly labelled. Documentation bundled with PSLab software does show not a few “how to connect instrument X” diagrams but unfortunately said diagrams picture a version 4 of the board and the current major version, V5, features radically different pinout (compare [3] with [4]/[5] and you will see immediately what I mean), not to mention that having to stare at a screen while wiring your circuit isn’t always optimal. Now, all versions of the board feature a complete set of header labels (along with LEDs showing the device is active) on the front side and at least the more recent ones additionally show more detailed descriptions on the back, clearly suggesting the optimal way to go is to make your case our of transparent material. But what if looking at the provided labels directly is not an option, for instance because you have gone eco-friendly and made your case out of wood? Probably stick a label to the front of the case… which brings us back to the problem of the case label from [5] not being compatible with recent versions of the board.

Which brings me to my take on adapting the design from [5] to match the header layout and labels of PSLab V5.1 as well as the laser-cut case design from [2]. It could probably be more accurate but having tried it out, it is close enough. Bluetooth and ICSP-programmer connectors near the centre of the board are not included because the current case design does not provide access to them and indeed, they haven’t even got headers soldered in. Licence and copyright: same as the original.

[1] https://pslab.io/

[2] https://github.com/fossasia/pslab-case

[3] https://github.com/fossasia/pslab-hardware/raw/master/docs/images/PSLab_v5_top.png

[4] https://github.com/fossasia/pslab-hardware/raw/master/docs/images/pslab_version_previews/PSLab_v4.png

[5] https://github.com/fossasia/pslab-hardware/raw/master/docs/images/pslabdesign.png

This is just a notification that the Blogs and the appropriate plug-ins for the release 5.1.1 have been updated.

With the release of these updated we (The Gentoo Blog Team) have updated the themes that had updates. If you have a blog on this site, and have a theme that is based on one of the following themes please consider updating as these themes are no longer updated and things will break in your blogs.

• KDE Breathe
• KDE Graffiti
• Oxygen
• The Following WordPress versions might stop working (Simply because of age)
  • Twenty Fourteen
  • Twenty Fifteen
  • Twenty Sixteen

If you are using one of these themes it is recommended that you update to the other themes available. If you think that there is an open source theme that you would like to have available please contact the Blogs team by opening a Bugzilla Bug with pertinent information.

We recently had to face free disk space outages on some of our scylla clusters and we learnt some very interesting things while outlining some improvements that could be made to the ScyllaDB guys.

100% disk space usage?

First of all I wanted to give a bit of a heads up about what happened when some of our scylla nodes reached (almost) 100% disk space usage.

Basically they:

• stopped listening to client requests
• complained in the logs
• wouldn’t flush commitlog (expected)
• abort their compaction work (which actually gave back a few GB of space)
• stay in a stuck / unable to stop state (unexpected, this has been reported)

After restarting your scylla server, the first and obvious thing you can try to do to get out of this situation is to run the nodetool clearsnapshot command which will remove any data snapshot that could be lying around. That’s a handy command to reclaim space usually.

Reminder: depending on your compaction strategy, it is usually not advised to allow your data to grow over 50% of disk space...

But that’s only a patch so let’s go down the rabbit hole and look at the optimization options we have.

Optimize your schemas

Schema design and the types your choose for your columns have a huge impact on disk space usage! And in our case we indeed overlooked some of the optimizations that we could have done from the start and that did cost us a lot of wasted disk space. Fortunately it was easy and fast to change.

To illustrate this, I’ll take a sample of 100,000 rows of a simple and naive schema associating readings of 50 integers to a user ID:

Note: all those operations were done using Scylla 3.0.3 on Gentoo Linux.

CREATE TABLE IF NOT EXISTS test.not_optimized
(
    uid text,
    readings list<int>,
    PRIMARY KEY(uid)
) WITH compression = {};

Once inserted on disk, this takes about 250MB of disk space:

250M    not_optimized-00cf1500520b11e9ae38000000000004

Now depending on your use case, if those readings at not meant to be updated for example you could use a frozen list instead, which will allow a huge storage optimization:

CREATE TABLE IF NOT EXISTS test.mid_optimized
 (
     uid text,
     readings frozen<list<int>>,
     PRIMARY KEY(uid)
 ) WITH compression = {};

With this frozen list we now consume 54MB of disk space for the same data!

54M     mid_optimized-011bae60520b11e9ae38000000000004

There’s another optimization that we could do since our user ID are UUIDs. Let’s switch to the uuid type instead of text:

CREATE TABLE IF NOT EXISTS test.optimized
 (
     uid uuid,
     readings frozen<list<int>>,
     PRIMARY KEY(uid)
 ) WITH compression = {};

By switching to uuid, we now consume 50MB of disk space: that’s a 80% reduced disk space consumption compared to the naive schema for the same data!

50M     optimized-01f74150520b11e9ae38000000000004

Enable compression

All those examples were not using compression. If your workload latencies allows it, you should probably enable compression on your sstables.

Let’s see its impact on our tables:

ALTER TABLE test.not_optimized WITH compression = {'sstable_compression': 'org.apache.cassandra.io.compress.LZ4Compressor'};
ALTER TABLE test.mid_optimized WITH compression = {'sstable_compression': 'org.apache.cassandra.io.compress.LZ4Compressor'};
ALTER TABLE test.optimized WITH compression = {'sstable_compression': 'org.apache.cassandra.io.compress.LZ4Compressor'};

Then we run a nodetool compact test to force a (re)compaction of all the sstables and we get:

63M     not_optimized-00cf1500520b11e9ae38000000000004
28M     mid_optimized-011bae60520b11e9ae38000000000004
24M     optimized-01f74150520b11e9ae38000000000004

Compression is really a great gain here allowing another 50% reduced disk space usage reduction on our optimized table!

Switch to the new “mc” sstable format

Since the Scylla 3.0 release you can use the latest “mc” sstable storage format on your scylla clusters. It promises a greater efficiency for usually a way more reduced disk space consumption!

It is not enabled by default, you have to add the enable_sstables_mc_format: true parameter to your scylla.yaml for it to be taken into account.

Since it’s backward compatible, you have nothing else to do as new compactions will start being made using the “mc” storage format and the scylla server will seamlessly read from old sstables as well.

But in our case of immediate disk space outage, we switched to the new format one node at a time, dropped the data from it and ran a nodetool rebuild to reconstruct the whole node using the new sstable format.

Let’s demonstrate its impact on our test tables: we add the option to the scylla.yaml file, restart scylla-server and run nodetool compact test again:

49M     not_optimized-00cf1500520b11e9ae38000000000004
26M     mid_optimized-011bae60520b11e9ae38000000000004
22M     optimized-01f74150520b11e9ae38000000000004

That’s a pretty cool gain of disk space, even more for the not optimized version of our schema!

So if you’re in great need of disk space or it is hard for you to change your schemas, switching to the new “mc” sstable format is a simple and efficient way to free up some space without effort.

Consider using secondary indexes

While denormalization is the norm (yep.. legitimate pun) in the NoSQL world this does not mean we have to duplicate everything all the time. A good example lies in the internals of secondary indexes if your workload can compromise with its moderate impact on latency.

Secondary indexes on scylla are built on top of Materialized Views that basically stores an up to date pointer from your indexed column to your main table partition key. That means that secondary indexes MVs are not duplicating all the columns (and thus the data) from your main table as you would have to do when denormalizing a table to query by another column: this saves disk space!

This of course comes with a latency drawback because if your workload is interested in the other columns than the partition key of the main table, the coordinator node will actually issue two queries to get all your data:

1. query the secondary index MV to get the pointer to the partition key of the main table
2. query the main table with the partition key to get the rest of the columns you asked for

This has been an effective trick to avoid duplicating a table and save disk space for some of our workloads!

(not a tip) Move the commitlog to another disk / partition?

This should only be considered as a sort of emergency procedure or for cost efficiency (cheap disk tiering) on non critical clusters.

While this is possible even if the disk is not formatted using XFS, it not advised to separate the commitlog from data on modern SSD/NVMe disks but… you technically can do it (as we did) on non production clusters.

Switching is simple, you just need to change the commitlog_directory parameter in your scylla.yaml file.

GNOME 3.30 is now available in Gentoo Linux testing branch.
Starting with this release, GNOME on Gentoo once again works with OpenRC, in addition to the usual systemd option. This is achieved through the elogind project, a standalone logind implementation based on systemd code, which is currently maintained by a fellow Gentoo user. It provides the missing logind interfaces currently required by GNOME without booting with systemd.

For easier GNOME install, the desktop/gnome profiles now set up default USE flags with elogind for OpenRC systems, while the desktop/gnome/systemd profiles continue to do that for systemd systems. Both have been updated to provide a better initial GNOME install experience. After profile selection, a full install should be simply a matter of `emerge gnome` for testing branch users. Don’t forget to adapt your system to any changed USE flags on previously installed packages too.

GNOME 3.32 is expected to be made available in testing branch soon as well, followed by introducing all this for stable branch users. This is hoped to complete within 6-8 weeks.

If you encounter issues, don’t hesitate to file bug reports or, if necessary, contact me via e-mail or IRC. You can also discuss the elogind aspects on the Gentoo Forums.

Acknowledgements

I’d like to thank Gavin Ferris, for kindly agreeing to sponsor my work on the above (upgrading GNOME on Gentoo from 3.26 to 3.30 and introducing Gentoo GNOME elogind support); and dantrell, for his pioneering overlay work integrating GNOME 3 with OpenRC on Gentoo, and also the GNOME and elogind projects.

I’m glad to announce a new (awaited) release of py3status featuring support for the sway window manager which allows py3status to enter the wayland environment!

Updated configuration and custom modules paths detection

The configuration section of the documentation explains the updated detection of the py3status configuration file (with respect of XDG_CONFIG environment variables):

• ~/.config/py3status/config
• ~/.config/i3status/config
• ~/.config/i3/i3status.conf
• ~/.i3status.conf
• ~/.i3/i3status.conf
• /etc/xdg/i3status/config
• /etc/i3status.conf

Regarding custom modules paths detection, py3status does as described in the documentation:

• ~/.config/py3status/modules
• ~/.config/i3status/py3status
• ~/.config/i3/py3status
• ~/.i3/py3status

Highlights

Lots of modules improvements and clean ups, see changelog.

• we worked on the documentation sections and content which allowed us to fix a bunch of typos
• our magic @lasers have worked a lot on harmonizing thresholds on modules along with a lot of code clean ups
• new module: scroll to scroll modules on your bar (#1748)
• @lasers has worked a lot on a more granular pango support for modules output (still work to do as it breaks some composites)

Thanks contributors

• Ajeet D’Souza
• @boucman
• Cody Hiar
• @cyriunx
• @duffydack
• @lasers
• Maxim Baz
• Thiago Kenji Okada
• Yaroslav Dronskii

I’m pretty sure that the title of this post will catch your attention…and/or maybe your curiosity.

Well..this is something I’m doing since years…and since did not cost too much to make it in a public and usable state, I decided to share my work, to help some people to avoid waste of time and to avoid to be angry when your cloud provider does not offer the gentoo image.

So what are the goals of this project?

1. Install gentoo on cloud providers that do not offer a Gentoo image (e.g Hetzner)
2. Install gentoo everywhere in few seconds.

To do a fast installation, we need a stage4….but what is exactly a stage4? In this case the stage4 is composed by the official gentoo stage3 plus grub, some more utilities and some file already configured.

So since the stage4 has already everything to complete the installation, we just need to make some replacement (fstab, grub and so on), install grub on the disk………..and…..it’s done (by the auto-installer script)!

At this point I’d expect some people to say….”yeah…it’s so simply and logical…why I didn’t think about that” – Well, I guess that every gentoo user didn’t discover that just after the first installation…so you don’t need to blame yourself

The technical details are covered by the README in the gentoo-stage4 git repository

As said in the README:

• If you have any request, feel free to contact me
• A star on the project will give me the idea of the usage and then the effort to put here.

So what’s more? Just a screenshot of the script in action

# Gentoo hetzner cloud
# Gentoo stage4
# Gentoo cloud

Traditionally, OpenPGP revocation certificates are used as a last resort. You are expected to generate one for your primary key and keep it in a secure location. If you ever lose the secret portion of the key and are unable to revoke it any other way, you import the revocation certificate and submit the updated key to keyservers. However, there is another interesting use for revocation certificates — revoking shared organization keys.

Let’s take Gentoo, for example. We are using a few keys needed to perform automated signatures on servers. For this reason, the key is especially exposed to attacks and we want to be able to revoke it quickly if the need arises. Now, we really do not want to have every single Infra member hold a copy of the secret primary key. However, we can give Infra members revocation certificates instead. This way, they maintain the possibility of revoking the key without unnecessarily increasing its exposure.

The problem with traditional revocation certificates is that they are supported for the purpose of revoking the primary key only. In our security model, the primary key is well protected, compared to subkeys that are totally exposed. Therefore, it is superfluous to revoke the complete key when only a subkey is compromised. To resolve this limitation, gen-revoke tool was created that can create exported revocation signatures for both the primary key and subkeys.

Technical background

The OpenPGP key (v4, as defined by RFC 4880) consists of a primary key, one or more UIDs and zero or more subkeys. Each of those keys and UIDs can include zero or more signature packets. Those packets bind information to the specific key or UID, and their authenticity is confirmed by a signature made using the secret portion of a primary key.

Signatures made by the key’s owner are called self-signatures. The most basic form of them serve as a binding between the primary key and its subkeys and UIDs. Since both those classes of objects are created independently of the primary key, self-signatures are necessary to distinguish authentic subkeys and UIDs created by the key owner from potential fakes. Appropriately, GnuPG will only accept subkeys and UIDs that have valid self-signature.

One specific type of signatures are revocation signatures. Those signatures indicate that the relevant key, subkey or UID has been revoked. If a revocation signature is found, it takes precedence over any other kinds of signatures and prevents the revoked object from being further used.

Key updates are means of distributing new data associated with the key. What’s important is that during an update the key is not replaced by a new one. Instead, GnuPG collects all the new data (subkeys, UIDs, signatures) and adds it to the local copy of the key. The validity of this data is verified against appropriate signatures. Appropriately, anyone can submit a key update to the keyserver, provided that the new data includes valid signatures. Similarly to local GnuPG instance, the keyserver is going to update its copy of the key rather than replacing it.

Revocation certificates specifically make use of this property. Technically, a revocation certificate is simply an exported form of a revocation signature, signed using the owner’s primary key. As long as it’s not on the key (i.e. GnuPG does not see it), it does not do anything. When it’s imported, GnuPG adds it to the key. Further submissions and exports include it, effectively distributing it to all copies of the key.

gen-revoke builds on this idea. It creates and exports revocation signatures for the primary key and subkeys. Due to implementation limitations (and for better compatibility), rather than exporting the signature alone it exports a minimal copy of the relevant key. This copy can be imported just like any other key export, and it causes the revocation signature to be added to the key. Afterwards, it can be exported and distributed just like a revocation done directly on the key.

Usage

To use the script, you need to have the secret portion of the primary key available, and public encryption keys for all the people who are supposed to obtain a copy of the revocation signatures (recipients).

The script takes at least two parameters: an identifier of the key for which revocation signatures should be created, followed by one or more e-mail addresses of signature recipients. It creates revocation signatures both for the primary key and for all valid subkeys, for all the people specified.

The signatures are written into the current directory as key exports and are encrypted to each specified person. They should be distributed afterwards, and kept securely by all the individuals. If a need to revoke either a subkey or the primary key arises, the first person available can decrypt the signature, import it and send the resulting key to keyservers.

Additionally, each signature includes a comment specifying the person it was created for. This comment will afterwards be displayed by GnuPG if one of the revocation signatures is imported. This provides a clear audit trace as to who revoked the key.

Security considerations

Each of the revocation signatures can be used by an attacker to disable the key in question. The signatures are protected through encryption. Therefore, the system is vulnerable to the key of a single signature owner being compromised.

However, this is considerably safer than the equivalent option of distributing the secret portion of the primary key. In the latter case, the attacker would be able to completely compromise the key and use it for malicious purposes; in the former, it is only capable of revoking the key and therefore causing some frustration. Furthermore, the revocation comment helps identifying the compromised user.

The tradeoff between reliability and security can be adjusted by changing the number of revocation signature holders.

This article describes the UI deficiency of Evolution mail client that extrapolates the trust of one of OpenPGP key UIDs into the key itself, and reports it along with the (potentially untrusted) primary UID. This creates the possibility of tricking the user into trusting a phished mail via adding a forged UID to a key that has a previously trusted UID.

Continue reading

Let’s say you want to send a confidential message to me, and possibly receive a reply. Through employing asymmetric encryption, you can prevent a third party from reading its contents, even if it can intercept the ciphertext. Through signatures, you can verify the authenticity of the message, and therefore detect any possible tampering. But for all this to work, you need to be able to verify the authenticity of the public keys first. In other words, we need to be able to prevent the aforementioned third party — possibly capable of intercepting your communications and publishing a forged key with my credentials on it — from tricking you into using the wrong key.

This renders key authenticity the fundamental problem of asymmetric cryptography. But before we start discussing how key certification is implemented, we need to cover another fundamental issue — identity. After all, who am I — who is the person you are writing to? Are you writing to a person you’ve met? Or to a specific Gentoo developer? Author of some project? Before you can distinguish my authentic key from a forged key, you need to be able to clearly distinguish me from an impostor.

Forms of identity

Identity via e-mail address

If your primary goal is to communicate with the owner of the particular e-mail address, it seems obvious to associate the identity with the owner of the e-mail address. However, how in reality would you distinguish a ‘rightful owner’ of the e-mail address from a cracker who managed to obtain access to it, or to intercept your network communications and inject forged mails?

The truth is, the best you can certify is that the owner of a particular key is able to read and/or send mails from a particular e-mail address, at a particular point in time. Then, if you can certify the same for a long enough period of time, you may reasonably assume the address is continuously used by the same identity (which may qualify as a legitimate owner or a cracker with a lot of patience).

Of course, all this relies on your trust in mail infrastructure not being compromised.

Identity via personal data

A stronger protection against crackers may be provided by associating the identity with personal data, as confirmed by government-issued documents. In case of OpenPGP, this is just the real name; X.509 certificates also provide fields for street address, phone number, etc.

The use of real names seems to be based on two assumptions: that your real name is reasonable well-known (e.g. it can be established with little risk of being replaced by a third party), and that the attacker does not wish to disclose his own name. Besides that, using real names meets with some additional criticism.

Firstly, requiring one to use his real name may be considered an invasion on privacy. Most notably, some people wish not to disclose or use their real names, and this effectively prevents them from ever being certified.

Secondly, real names are not unique. After all, the naming systems developed from the necessity of distinguishing individuals in comparatively small groups, and they simply don’t scale to the size of the Internet. Therefore, name collisions are entirely possible and we are relying on sheer luck that the attacker wouldn’t happen to have the same name as you do.

Thirdly and most importantly, verifying identity documents is non-trivial and untrained individuals are likely to fall victim of mediocre quality fakes. After all, we’re talking about people who hopefully read some article on verifying a particular kind of document but have no experience recognizing forgery, no specialized hardware (I suppose most of you don’t carry a magnifying glass and a UV light on yourself) and who may lack skills in comparing signatures or photographs (not to mention some people have really old photographs in documents). Some countries don’t even issue any official documentation for document verification in English!

Finally, even besides the point of forged documents, this relies on trust in administration.

Identity via photographs

This one I’m mentioning merely for completeness. OpenPGP keys allow adding a photo as one of your UIDs. However, this is rather rarely used (out of the keys my GnuPG fetched so far, less than 10% have photographs). The concerns are similar as for personal data: it assumes that others are reliably able to know how you look like, and that they are capable of reliably comparing faces.

Online identity

An interesting concept is to use your public online activity to prove your identity — such as websites or social media. This is generally based on cross-referencing multiple resources with cryptographically proven publishing access, and assuming that an attacker would not be able to compromise all of them simultaneously.

A form of this concept is utilized by keybase.io. This service builds trust in user profiles via cryptographically cross-linking your profiles on some external sites and/or your websites. Furthermore, it actively encourages other users to verify those external proofs as well.

This identity model entirely relies on trust in network infrastructure and external sites. The likeliness of it being compromised is reduced by (potentially) relying on multiple independent sites.

Web of Trust model

Most of time, you won’t be able to directly verify the identity of everyone you’d like to communicate with. This creates a necessity of obtaining indirect proof of authenticity, and the model normally used for that purpose in OpenPGP is the Web of Trust. I won’t be getting into the fine details — you can find them e.g. in the GNU Privacy Handbook. For our purposes, it suffices to say that in WoT the authenticity of keys you haven’t verified may be assessed by people whose keys you trust already, or people they know, with a limited level of recursion.

The more key holders you can trust, the more keys you can have verified indirectly and the more likely it is that your future recipient will be in that group. Or that you will be able to get someone from across the world into your WoT by meeting someone residing much closer to yourself. Therefore, you’d naturally want the WoT to grow fast and include more individuals. You’d want to preach OpenPGP onto non-crypto-aware people. However, this comes with inherent danger: can you really trust that they will properly verify the identity of the keys they sign?

I believe this is the most fundamental issue with WoT model: for it to work outside of small specialized circles, it has to include more and more individuals across the world. But this growth inevitable makes it easier for a malicious third party to find people that can be tricked into certifying keys with forged identities.

Conclusion

The fundamental problem in OpenPGP usage is finding the correct key and verifying its authenticity. This becomes especially complex given there is no single clear way of determining one’s identity in the Internet. Normally, OpenPGP uses a combination of real name and e-mail address, optionally combined with a photograph. However, all of them have their weaknesses.

Direct identity verification for all recipients is non-practical, and therefore requires indirect certification solutions. While the WoT model used by OpenPGP attempts to avoid centralized trust specific to PKI, it is not clear whether it’s practically manageable. On one hand, it requires trusting more people in order to improve coverage; on the other, it makes it more vulnerable to fraud.

Given all the above, the trust-via-online-presence concept may be of some interest. Most importantly, it establishes a closer relationship between the identity you actually need and the identity you verify — e.g. you want to mail the person being an open source developer, author of some specific projects rather than arbitrary person with a common enough name. However, this concept is not established broadly yet.

This article shortly explains the historical git weakness regarding handling commits with multiple OpenPGP signatures in git older than v2.20. The method of creating such commits is presented, and the results of using them are described and analyzed.

Continue reading

Two py3status versions in less than a month? That’s the holidays effect but not only!

Our community has been busy discussing our way forward to 4.0 (see below) and organization so it was time I wrote a bit about that.

Community

A new collaborator

First of all we have the great pleasure and honor to welcome Maxim Baz @maximbaz as a new collaborator on the project!

His engagement, numerous contributions and insightful reviews to py3status has made him a well known community member, not to mention his IRC support

Once again, thank you for being there Maxim!

Zen of py3status

As a result of an interesting discussion, we worked on defining better how to contribute to py3status as well as a set of guidelines we agree on to get the project moving on smoothly.

Here is born the zen of py3status which extends the philosophy from the user point of view to the contributor point of view!

This allowed us to handle the numerous open pull requests and get their number down to 5 at the time of writing this post!

Even our dear @lasers don’t have any open PR anymore

3.15 + 3.16 versions

Our magic @lasers has worked a lot on general modules options as well as adding support for i3-gaps added features such as border coloring and fine tuning.

Also interesting is the work of Thiago Kenji Okada @m45t3r around NixOS packaging of py3status. Thanks a lot for this work and for sharing Thiago!

I also liked the question of Andreas Lundblad @aioobe asking if we could have a feature allowing to display a custom graphical output, such as a small PNG or anything upon clicking on the i3bar, you might be interested in following up the i3 issue he opened.

Make sure to read the amazing changelog for details, a lot of modules have been enhanced!

Highlights

• You can now set a background, border colors and their urgent counterparts on a global scale or per module
• CI now checks for black format on modules, so now all the code base obey the black format style!
• All HTTP requests based modules now have a standard way to define HTTP timeout as well as the same 10 seconds default timeout
• py3-cmd now allows sending click events with modifiers
• The py3status -n / –interval command line argument has been removed as it was obsolete. We will ignore it if you have set it up, but better remove it to be clean
• You can specify your own i3status binary path using the new -u, –i3status command line argument thanks to @Dettorer and @lasers
• Since Yahoo! decided to retire its public & free weather API, the weather_yahoo module has been removed

New modules

• new conky module: display conky system monitoring (#1664), by lasers
• new module emerge_status: display information about running gentoo emerge (#1275), by AnwariasEu
• new module hueshift: change your screen color temperature (#1142), by lasers
• new module mega_sync: to check for MEGA service synchronization (#1458), by Maxim Baz
• new module speedtest: to check your internet bandwidth (#1435), by cyrinux
• new module usbguard: control usbguard from your bar (#1376), by cyrinux
• new module velib_metropole: display velib metropole stations and (e)bikes (#1515), by cyrinux

A word on 4.0

Do you wonder what’s gonna be in the 4.0 release?
Do you have ideas that you’d like to share?
Do you have dreams that you’d love to become true?

Then make sure to read and participate in the open RFC on 4.0 version!

Development has not started yet; we really want to hear from you.

Thank you contributors!

There would be no py3status release without our amazing contributors, so thank you guys!

• AnwariasEu
• cyrinux
• Dettorer
• ecks
• flyingapfopenguin
• girst
• Jack Doan
• justin j lin
• Keith Hughitt
• L0ric0
• lasers
• Maxim Baz
• oceyral
• Simon Legner
• sridhars
• Thiago Kenji Okada
• Thomas F. Duellmann
• Till Backhaus

It’s been almost one month since I had the chance to attend and speak at Scylla Summit 2018 so I’m relieved to finally publish a short write-up on the key things I wanted to share about this wonderful event!

Make Scylla boring

This statement of Glauber Costa sums up what looked to me to be the main driver of the engineering efforts put into Scylla lately: making it work so consistently well on any kind of workload that it’s boring to operate

I will follow up on this statement to highlight the things I heard and (hopefully) understood during the summit. I hope you’ll find it insightful.

Reduced operational efforts

The thread-per-core and queues design still has a lot of possibilities to be leveraged.

The recent addition of RPC streaming capabilities to seastar allows a drastic reduction in the time it takes the cluster to grow or shrink (data rebalancing / resynchronization).

Incremental compaction is also very promising as this background process is one of the most expensive there is in the database’s design.

I was happy to hear that scylla-manager will soon be made available and free to use with basic features while retaining more advanced ones for enterprise version (like backup/restore).
I also noticed that the current version was not supporting SSL enabled clusters to store its configuration. So I directly asked Michał for it and I’m glad that it will be released on version 1.3.1.

Performant multi-tenancy

Why choose between real-time OLTP & analytics OLAP workloads?

The goal here is to be able to run both on the same cluster by giving users the ability to assign “SLA” shares to ROLES. That’s basically like pools on Hadoop at a much finer grain since it will create dedicated queues that will be weighted by their share.

Having one queue per usage and full accounting will allow to limit resources efficiently and users to have their say on their latency SLAs.

But Scylla also has a lot to do in the background to run smoothly. So while this design pattern was already applied to tamper compactions, a lot of work has also been done on automatic flow control and back pressure.

For instance, Materialized Views are updated asynchronously which means that while we can interact and put a lot of pressure on the table its based on (called the Main Table), we could overwhelm the background work that’s needed to keep MVs View Tables in sync. To mitigate this, a smart back pressure approach was developed and will throttle the clients to make sure that Scylla can manage to do everything at the best performance the hardware allows!

I was happy to hear that work on tiered storage is also planned to better optimize disk space costs for certain workloads.

Last but not least, columnar storage optimized for time series and analytics workloads are also something the developers are looking at.

Latency is expensive

If you care for latency, you might be happy to hear that a new polling API (named IOCB_CMD_POLL) has been contributed by Christoph Hellwig and Avi Kivity to the 4.19 Linux kernel which avoids context switching I/O by using a shared ring between kernel and userspace. Scylla will be using it by default if the kernel supports it.

The iotune utility has been upgraded since 2.3 to generate an enhanced I/O configuration.

Also, persistent (disk backed) in-memory tables are getting ready and are very promising for latency sensitive workloads!

A word on drivers

ScyllaDB has been relying on the Datastax drivers since the start. While it’s a good thing for the whole community, it’s important to note that the shard-per-CPU approach on data that Scylla is using is not known and leveraged by the current drivers.

Discussions took place and it seems that Datastax will not allow the protocol to evolve so that drivers could discover if the connected cluster is shard aware or not and then use this information to be more clever in which write/read path to use.

So for now ScyllaDB has been forking and developing their shard aware drivers for Java and Go (no Python yet… I was disappointed).

Kubernetes & containers

The ScyllaDB guys of course couldn’t avoid the Kubernetes frenzy so Moreno Garcia gave a lot of feedback and tips on how to operate Scylla on docker with minimal performance degradation.

Kubernetes has been designed for stateless applications, not stateful ones and Docker does some automatic magic that have rather big performance hits on Scylla. You will basically have to play with affinities to dedicate one Scylla instance to run on one server with a “retain” reclaim policy.

Remember that the official Scylla docker image runs with dev-mode enabled by default which turns off all performance checks on start. So start by disabling that and look at all the tips and literature that Moreno has put online!

Scylla 3.0

A lot has been written on it already so I will just be short on things that important to understand in my point of view.

• Materialized Views do back fill the whole data set
  
  • this job is done by the view building process
  • you can watch its progress in the system_distributed.view_build_status table
    
• Secondary Indexes are Materialized Views under the hood
  
  • it’s like a reverse pointer to the primary key of the Main Table
    
  • so if you read the whole row by selecting on the indexed column, two reads will be issued under the hood: one on the indexed MV view table to get the primary key and one on the main table to get the rest of the columns
    
  • so if your workload is mostly interested by the whole row, you’re better off creating a complete MV to read from than using a SI
  • this is even more true if you plan to do range scans as this double query could lead you to read from multiple nodes instead of one
    
• Range scan is way more performant
  • ALLOW FILTERING finally allows a great flexibility by providing server-side filtering!

Random notes

Support for LWT (lightweight transactions) will be relying on a future implementation of the Raft consensus algorithm inside Scylla. This work will also benefits Materialized Views consistency. Duarte Nunes will be the one working on this and I envy him very much!

Support for search workloads is high in the ScyllaDB devs priorities so we should definitely hear about it in the coming months.

Support for “mc” sstables (new generation format) is done and will reduce storage requirements thanks to metadata / data compression. Migration will be transparent because Scylla can read previous formats as well so it will upgrade your sstables as it compacts them.

ScyllaDB developers have not settled on how to best implement CDC. I hope they do rather soon because it is crucial in their ability to integrate well with Kafka!

Materialized Views, Secondary Indexes and filtering will benefit from the work on partition key and indexes intersections to avoid server side filtering on the coordinator. That’s an important optimization to come!

Last but not least, I’ve had the pleasure to discuss with Takuya Asada who is the packager of Scylla for RedHat/CentOS & Debian/Ubuntu. We discussed Gentoo Linux packaging requirements as well as the recent and promising work on a relocatable package. We will collaborate more closely in the future!

The tar format is one of the oldest archive formats in use. It comes as no surprise that it is ugly — built as layers of hacks on the older format versions to overcome their limitations. However, given the POSIX standarization in late 80s and the popularity of GNU tar, you would expect the interoperability problems to be mostly resolved nowadays.

This article is directly inspired by my proof-of-concept work on new binary package format for Gentoo. My original proposal used volume label to provide user- and file(1)-friendly way of distinguish our binary packages. While it is a GNU tar extension, it falls within POSIX ustar implementation-defined file format and you would expect that non-compliant implementations would extract it as regular files. What I did not anticipate is that some implementation reject the whole archive instead.

This naturally raised more questions on how portable various tar formats actually are. To verify that, I have decided to analyze the standards for possible incompatibility dangers and build a suite of test inputs that could be used to check how various implementations cope with that. This article describes those points and provides test results for a number of implementations.

Please note that this article is focused merely on read-wise format compatibility. In other words, it establishes how tar files should be written in order to achieve best probability that it will be read correctly afterwards. It does not investigate what formats the listed tools can write and whether they can correctly create archives using specific features.

Continue reading

I’m happy to announce this release as it contains some very interesting developments in the project. This release was focused on core changes.

IMPORTANT notice

There are now two optional dependencies to py3status:

• gevent
  • will monkey patch the code to make it concurrent
  • the main benefit is to use an asynchronous loop instead of threads
    
• pyudev
  • will enable a udev monitor if a module asks for it (only xrandr so far)
  • the benefit is described below
    

To install them all using pip, simply do:

pip install py3status[all]

Modules can now react/refresh on udev events

When pyudev is available, py3status will allow modules to subscribe and react to udev events!

The xrandr module uses this feature by default which allows the module to instantly refresh when you plug in or off a secondary monitor. This also allows to stop running the xrandr command in the background and saves a lot of CPU!

Highlights

• py3status core uses black formatter
• fix default i3status.conf detection
  • add ~/.config/i3 as a default config directory, closes #1548
  • add .config/i3/py3status in default user modules include directories
• add markup (pango) support for modules (#1408), by @MikaYuoadas
• py3: notify_user module name in the title (#1556), by @lasers
• print module information to sdtout instead of stderr (#1565), by @robertnf
• battery_level module: default to using sys instead of acpi (#1562), by @eddie-dunn
• imap module: fix output formatting issue (#1559), by @girst

Thank you contributors!

• eddie-dunn
• girst
• MikaYuoadas
• robertnf
• lasers
• maximbaz
• tobes

I am once again lagging behind the release blog posts but this one is an important one.

I’m proud to announce that our long time contributor @lasers has become an official collaborator of the py3status project!

Dear @lasers, your amazing energy and overwhelming ideas have served our little community for a while. I’m sure we’ll have a great way forward as we learn to work together with @tobes Thank you again very much for everything you do!

This release is as much dedicated to you as it is yours

IMPORTANT notice

After this release, py3status coding style CI will enforce the ‘black‘ formatter style.

Highlights

Needless to say that the changelog is huge, as usual, here is a very condensed view:

• documentation updates, especially on the formatter (thanks @L0ric0)
  
• py3 storage: use $XDG_CACHE_HOME or ~/.cache
• formatter: multiple variable and feature fixes and enhancements
• better config parser
• new modules: lm_sensors, loadavg, mail, nvidia_smi, sql, timewarrior, wanda_the_fish

Thank you contributors!

• lasers
• tobes
• maximbaz
• cyrinux
• Lorenz Steinert @L0ric0
• wojtex
• horgix
• su8
• Maikel Punie

On 2018-09-15 meeting, the Trustees have given the final stamp of approval to the new Gentoo copyright policy outlined in GLEP 76. This policy is the result of work that has been slowly progressing since 2005, and that has taken considerable speed by the end of 2017. It is a major step forward from the status quo that has been used since the forming of Gentoo Foundation, and that mostly has been inherited from earlier Gentoo Technologies.

The policy aims to cover all copyright-related aspects, bringing Gentoo in line with the practices used in many other large open source projects. Most notably, it introduces a concept of Gentoo Certificate of Origin that requires all contributors to confirm that they are entitled to submit their contributions to Gentoo, and corrects the copyright attribution policy to be viable under more jurisdictions.

This article aims to shortly reiterate over the most important points in the new copyright policy, and provide a detailed guide on following it in Q&A form.

Continue reading

With Qt5 gaining support for high-DPI displays, and applications starting to exercise that support, it’s easy for applications to suddenly become unusable with some screens. For example, my old Samsung TV reported itself as 7″ screen. While this used not to really matter with websites forcing you to force the resolution of 96 DPI, the high-DPI applications started scaling themselves to occupy most of my screen, with elements becoming really huge (and ugly, apparently due to some poor scaling).

It turns out that it is really hard to find a solution for this. Most of the guides and tips are focused either on proprietary drivers or on getting custom resolutions. The DisplaySize specification in xorg.conf apparently did not change anything either. Finally, I was able to resolve the issue by overriding the EDID data for my screen. This guide explains how I did it.

Step 1: dump EDID data

Firstly, you need to get the EDID data from your monitor. Supposedly read-edid tool could be used for this purpose but it did not work for me. With only a little bit more effort, you can get it e.g. from xrandr:

$ xrandr --verbose
[...]
HDMI-0 connected primary 1920x1080+0+0 (0x57) normal (normal left inverted right x axis y axis) 708mm x 398mm
[...]
  EDID:
    00ffffffffffff004c2dfb0400000000
    2f120103804728780aee91a3544c9926
    0f5054bdef80714f8100814081809500
    950fb300a940023a801871382d40582c
    4500c48e2100001e662150b051001b30
    40703600c48e2100001e000000fd0018
    4b1a5117000a2020202020200000000a
    0053414d53554e470a20202020200143
    020323f14b901f041305140312202122
    2309070783010000e2000f67030c0010
    00b82d011d007251d01e206e285500c4
    8e2100001e011d00bc52d01e20b82855
    40c48e2100001e011d8018711c162058
    2c2500c48e2100009e011d80d0721c16
    20102c2580c48e2100009e0000000000
    00000000000000000000000000000029
[...]

If you have multiple displays connected, make sure to use the EDID for the one you’re overriding. Copy the hexdump and convert it to a binary blob. You can do this by passing it through xxd -p -r (installed by vim).

Step 2: fix screen dimensions

Once you have the EDID blob ready, you need to update the screen dimensions inside it. Initially, I did it using hex editor which involved finding all the occurrences, updating them (and manually encoding into the weird split-integers) and correcting the checksums. Then, I’ve written edid-fixdim so you wouldn’t have to repeat that experience.

First, use --get option to verify that your EDID is supported correctly:

$ edid-fixdim -g edid.bin
EDID structure: 71 cm x 40 cm
Detailed timing desc: 708 mm x 398 mm
Detailed timing desc: 708 mm x 398 mm
CEA EDID found
Detailed timing desc: 708 mm x 398 mm
Detailed timing desc: 708 mm x 398 mm
Detailed timing desc: 708 mm x 398 mm
Detailed timing desc: 708 mm x 398 mm

So your EDID consists of basic EDID structure, followed by one extension block. The screen dimensions are stored in 7 different blocks you’d have to update, and referenced in two checksums. The tool will take care of updating it all for you, so just pass the correct dimensions to --set:

$ edid-fixdim -s 1600x900 edid.bin
EDID structure updated to 160 cm x 90 cm
Detailed timing desc updated to 1600 mm x 900 mm
Detailed timing desc updated to 1600 mm x 900 mm
CEA EDID found
Detailed timing desc updated to 1600 mm x 900 mm
Detailed timing desc updated to 1600 mm x 900 mm
Detailed timing desc updated to 1600 mm x 900 mm
Detailed timing desc updated to 1600 mm x 900 mm

Afterwards, you can use --get again to verify that the changes were made correctly.

Step 3: overriding EDID data

Now it’s just the matter of putting the override in motion. First, make sure to enable CONFIG_DRM_LOAD_EDID_FIRMWARE in your kernel:

Device Drivers  --->
  Graphics support  --->
    Direct Rendering Manager (XFree86 4.1.0 and higher DRI support)  --->
      [*] Allow to specify an EDID data set instead of probing for it

Then, determine the correct connector name. You can find it in dmesg output:

$ dmesg | grep -C 1 Connector
[   15.192088] [drm] ib test on ring 5 succeeded
[   15.193461] [drm] Radeon Display Connectors
[   15.193524] [drm] Connector 0:
[   15.193580] [drm]   HDMI-A-1
--
[   15.193800] [drm]     DFP1: INTERNAL_UNIPHY1
[   15.193857] [drm] Connector 1:
[   15.193911] [drm]   DVI-I-1
--
[   15.194210] [drm]     CRT1: INTERNAL_KLDSCP_DAC1
[   15.194267] [drm] Connector 2:
[   15.194322] [drm]   VGA-1

Copy the new EDID blob into location of your choice inside /lib/firmware:

$ mkdir /lib/firmware/edid
$ cp edid.bin /lib/firmware/edid/samsung.bin

Finally, add the override to your kernel command-line:

drm.edid_firmware=HDMI-A-1:edid/samsung.bin

If everything went fine, xrandr should report correct screen dimensions after next reboot, and dmesg should report that EDID override has been loaded:

$ dmesg | grep EDID
[   15.549063] [drm] Got external EDID base block and 1 extension from "edid/samsung.bin" for connector "HDMI-A-1"

If it didn't, check dmesg for error messages.

I have recently worked on enabling 2-step authentication via SSH on the Gentoo developer machine. I have selected google-authenticator-libpam amongst different available implementations as it seemed the best maintained and having all the necessary features, including a friendly tool for users to configure it. However, its design has a weakness: it stores the secret unprotected in user’s home directory.

This means that if an attacker manages to gain at least temporary access to the filesystem with user’s privileges — through a malicious process, vulnerability or simply because someone left the computer unattended for a minute — he can trivially read the secret and therefore clone the token source without leaving a trace. It would completely defeat the purpose of the second step, and the user may not even notice until the attacker makes real use of the stolen secret.

In order to protect against this, I’ve created google-authenticator-wrappers (as upstream decided to ignore the problem). This package provides a rather trivial setuid wrapper that manages a write-only, authentication-protected secret store for the PAM module. Additionally, it comes with a test program (so you can test the OTP setup without jumping through the hoops or risking losing access) and friendly wrappers for the default setup, as used on Gentoo Infra.

The recommended setup (as utilized by sys-auth/google-authenticator-wrappers package) is to use a dedicated user for the password store. In this scenario, the users are unable to read their secrets, and all secret operations (including authentication via the PAM module) are done using an unprivileged user. Furthermore, any operation regarding the configuration (either updating it or removing the second step) require regular PAM authentication (e.g. typing your own password).

This is consistent with e.g. how shadow operates (users can’t read their passwords, nor update them without authenticating first), how most sites using 2-factor authentication operate (again, users can’t read their secrets) and follows the RFC 6238 recommendation (that keys […] SHOULD be protected against unauthorized access and usage). It solves the aforementioned issue by preventing user-privileged processes from reading the secrets and recovery codes. Furthermore, it prevents the attacker with this particular level of access from disabling 2-step authentication, changing the secret or even weakening the configuration.

Integricloud gave me access to their infrastructure to track some issues on ppc64 and ppc64le.

Since some of the issues are related to the compilers, I obviously installed Gentoo on it and in the process I started to fix some issues with catalyst to get a working install media, but that’s for another blogpost.

Today I’m just giving a walk-through on how to get a ppc64le (and ppc64 soon) VM up and running.

Preparation

Read this and get your install media available to your instance.

Install Media

I’m using the Gentoo installcd I’m currently refining.

Booting

You have to append console=hvc0 to your boot command, the boot process might figure it out for you on newer install medias (I still have to send patches to update livecd-tools)

Network configuration

You have to manually setup the network.
You can use ifconfig and route or ip as you like, refer to your instance setup for the parameters.

ifconfig enp0s0 ${ip}/16
route add -net default gw ${gw}
echo "nameserver 8.8.8.8" > /etc/resolv.conf

ip a add ${ip}/16 dev enp0s0
ip l set enp0s0 up
ip r add default via ${gw}
echo "nameserver 8.8.8.8" > /etc/resolv.conf

Disk Setup

OpenFirmware seems to like gpt much better:

parted /dev/sda mklabel gpt

You may use fdisk to create:
– a PowerPC PrEP boot partition of 8M
– root partition with the remaining space

Device     Start      End  Sectors Size Type
/dev/sda1   2048    18431    16384   8M PowerPC PReP boot
/dev/sda2  18432 33554654 33536223  16G Linux filesystem

I’m using btrfs and zstd-compress /usr/portage and /usr/src/.

mkfs.btrfs /dev/sda2

Initial setup

It is pretty much the usual.

mount /dev/sda2 /mnt/gentoo
cd /mnt/gentoo
wget https://dev.gentoo.org/~mattst88/ppc-stages/stage3-ppc64le-20180810.tar.xz
tar -xpf stage3-ppc64le-20180810.tar.xz
mount -o bind /dev dev
mount -t devpts devpts dev/pts
mount -t proc proc proc
mount -t sysfs sys sys
cp /etc/resolv.conf etc
chroot .

You just have to emerge grub and gentoo-sources, I diverge from the defconfig by making btrfs builtin.

My /etc/portage/make.conf:

CFLAGS="-O3 -mcpu=power9 -pipe"
# WARNING: Changing your CHOST is not something that should be done lightly.
# Please consult https://wiki.gentoo.org/wiki/Changing_the_CHOST_variable beforee
 changing.
CHOST="powerpc64le-unknown-linux-gnu"

# NOTE: This stage was built with the bindist Use flag enabled
PORTDIR="/usr/portage"
DISTDIR="/usr/portage/distfiles"
PKGDIR="/usr/portage/packages"

USE="ibm altivec vsx"

# This sets the language of build output to English.
# Please keep this setting intact when reporting bugs.
LC_MESSAGES=C
ACCEPT_KEYWORDS=~ppc64

MAKEOPTS="-j4 -l6"
EMERGE_DEFAULT_OPTS="--jobs 10 --load-average 6 "

My minimal set of packages I need before booting:

emerge grub gentoo-sources vim btrfs-progs openssh

NOTE: You want to emerge again openssh and make sure bindist is not in your USE.

Kernel & Bootloader

cd /usr/src/linux
make defconfig
make menuconfig # I want btrfs builtin so I can avoid a initrd
make -j 10 all && make install && make modules_install
grub-install /dev/sda1
grub-mkconfig -o /boot/grub/grub.cfg

NOTE: make sure you pass /dev/sda1 otherwise grub will happily assume OpenFirmware knows about btrfs and just point it to your directory.
That’s not the case unfortunately.

Networking

I’m using netifrc and I’m using the eth0-naming-convention.

touch /etc/udev/rules.d/80-net-name-slot.rules
ln -sf /etc/init.d/net.{lo,eth0}
echo -e "config_eth0=\"${ip}/16\"\nroutes_eth0="default via ${gw}\"\ndns_servers_eth0=\"8.8.8.8\"" > /etc/conf.d/net

Password and SSH

Even if the mticlient is quite nice, you would rather use ssh as much as you could.

passwd 
rc-update add sshd default

Finishing touches

Right now sysvinit does not add the hvc0 console as it should due to a profile quirk, for now check /etc/inittab and in case add:

echo 'hvc0:2345:respawn:/sbin/agetty -L 9600 hvc0' >> /etc/inittab

Add your user and add your ssh key and you are ready to use your new system!

This quick article is a wrap up for reference on how to connect to ScyllaDB using Spark 2 when authentication and SSL are enforced for the clients on the Scylla cluster.

We encountered multiple problems, even more since we distribute our workload using a YARN cluster so that our worker nodes should have everything they need to connect properly to Scylla.

We found very little help online so I hope it will serve anyone facing similar issues (that’s also why I copy/pasted them here).

The authentication part is easy going by itself and was not the source of our problems, SSL on the client side was.

Environment

• (py)spark: 2.1.0.cloudera2
• spark-cassandra-connector: datastax:spark-cassandra-connector: 2.0.1-s_2.11
• python: 3.5.5
• java: 1.8.0_144
• scylladb: 2.1.5

SSL cipher setup

The Datastax spark cassandra driver uses default the TLS_RSA_WITH_AES_256_CBC_SHA cipher that the JVM does not support by default. This raises the following error when connecting to Scylla:

18/07/18 13:13:41 WARN channel.ChannelInitializer: Failed to initialize a channel. Closing: [id: 0x8d6f78a7]
java.lang.IllegalArgumentException: Cannot support TLS_RSA_WITH_AES_256_CBC_SHA with currently installed providers

According to the ssl documentation we have two ciphers available:

1. TLS_RSA_WITH_AES_256_CBC_SHA
2. TLS_RSA_WITH_AES_128_CBC_SHA

We can get get rid of the error by lowering the cipher to TLS_RSA_WITH_AES_128_CBC_SHA using the following configuration:

.config("spark.cassandra.connection.ssl.enabledAlgorithms", "TLS_RSA_WITH_AES_128_CBC_SHA")\

However, this is not really a good solution and instead we’d be inclined to use the TLS_RSA_WITH_AES_256_CBC_SHA version. For this we need to follow this Datastax’s procedure.

Then we need to deploy the JCE security jars on our all client nodes, if using YARN like us this means that you have to deploy these jars to all your NodeManager nodes.

For example by hand:

# unzip jce_policy-8.zip
# cp UnlimitedJCEPolicyJDK8/*.jar /opt/oracle-jdk-bin-1.8.0.144/jre/lib/security/

Java trust store

When connecting, the clients need to be able to validate the Scylla cluster’s self-signed CA. This is done by setting up a trustStore JKS file and providing it to the spark connector configuration (note that you protect this file with a password).

keyStore vs trustStore

In SSL handshake purpose of trustStore is to verify credentials and purpose of keyStore is to provide credentials. keyStore in Java stores private key and certificates corresponding to the public keys and is required if you are a SSL Server or SSL requires client authentication. TrustStore stores certificates from third parties or your own self-signed certificates, your application identify and validates them using this trustStore.

The spark-cassandra-connector documentation has two options to handle keyStore and trustStore.

When we did not use the trustStore option, we would get some obscure error when connecting to Scylla:

com.datastax.driver.core.exceptions.TransportException: [node/1.1.1.1:9042] Channel has been closed

When enabling DEBUG logging, we get a clearer error which indicated a failure in validating the SSL certificate provided by the Scylla server node:

Caused by: sun.security.validator.ValidatorException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target

setting up the trustStore JKS

You need to have the self-signed CA public certificate file, then issue the following command:

# keytool -importcert -file /usr/local/share/ca-certificates/MY_SELF_SIGNED_CA.crt -keystore COMPANY_TRUSTSTORE.jks -noprompt
Enter keystore password:  
Re-enter new password: 
Certificate was added to keystore

using the trustStore

Now you need to configure spark to use the trustStore like this:

.config("spark.cassandra.connection.ssl.trustStore.password", "PASSWORD")\
.config("spark.cassandra.connection.ssl.trustStore.path", "COMPANY_TRUSTSTORE.jks")\

Spark SSL configuration example

This wraps up the SSL connection configuration used for spark.

This example uses pyspark2 and reads a table in Scylla from a YARN cluster:

$ pyspark2 --packages datastax:spark-cassandra-connector:2.0.1-s_2.11 --files COMPANY_TRUSTSTORE.jks

>>> spark = SparkSession.builder.appName("scylla_app")\
.config("spark.cassandra.auth.password", "test")\
.config("spark.cassandra.auth.username", "test")\
.config("spark.cassandra.connection.host", "node1,node2,node3")\
.config("spark.cassandra.connection.ssl.clientAuth.enabled", True)\
.config("spark.cassandra.connection.ssl.enabled", True)\
.config("spark.cassandra.connection.ssl.trustStore.password", "PASSWORD")\
.config("spark.cassandra.connection.ssl.trustStore.path", "COMPANY_TRUSTSTORE.jks")\
.config("spark.cassandra.input.split.size_in_mb", 1)\
.config("spark.yarn.queue", "scylla_queue").getOrCreate()

>>> df = spark.read.format("org.apache.spark.sql.cassandra").options(table="my_table", keyspace="test").load()
>>> df.show()

I felt like sharing a recent story that allowed us identify a bot in a haystack thanks to Scylla.

The scenario

While working on loading up 2B+ of rows into Scylla from Hive (using Spark), we noticed a strange behaviour in the performances of one of our nodes:

So we started wondering why that server in blue was having those peaks of load and was clearly diverging from the two others… As we obviously expect the three nodes to behave the same, there were two options on the table:

1. hardware problem on the node
2. bad data distribution (bad schema design? consistent hash problem?)

We shared this with our pals from ScyllaDB and started working on finding out what was going on.

The investigation

Hardware?

Hardware problem was pretty quickly evicted, nothing showed up on the monitoring and on the kernel logs. I/O queues and throughput were good:

Data distribution?

Avi Kivity (ScyllaDB’s CTO) quickly got the feeling that something was wrong with the data distribution and that we could be facing a hotspot situation. He quickly nailed it down to shard 44 thanks to the scylla-grafana-monitoring platform.

Data is distributed between shards that are stored on nodes (consistent hash ring). This distribution is done by hashing the primary key of your data which dictates the shard it belongs to (and thus the node(s) where the shard is stored).

If one of your keys is over represented in your original data set, then the shard it belongs to can be overly populated and the related node overloaded. This is called a hotspot situation.

tracing queries

The first step was to trace queries in Scylla to try to get deeper into the hotspot analysis. So we enabled tracing using the following formula to get about 1 trace per second in the system_traces namespace.

tracing probability = 1 / expected requests per second throughput

In our case, we were doing between 90K req/s and 150K req/s so we settled for 100K req/s to be safe and enabled tracing on our nodes like this:

# nodetool settraceprobability 0.00001

Turns out tracing didn’t help very much in our case because the traces do not include the query parameters in Scylla 2.1, it is becoming available in the soon to be released 2.2 version.

NOTE: traces expire on the tables, make sure your TRUNCATE the events and sessions tables while iterating. Else you will have to wait for the next gc_grace_period (10 days by default) before they are actually removed. If you do not do that and generate millions of traces like we did, querying the mentioned tables will likely time out because of the “tombstoned” rows even if there is no trace inside any more.

looking at cfhistograms

Glauber Costa was also helping on the case and got us looking at the cfhistograms of the tables we were pushing data to. That proved to be clearly highlighting a hotspot problem:

histograms
Percentile  SSTables     Write Latency      Read Latency    Partition Size        Cell Count
                             (micros)          (micros)           (bytes)                  
50%             0,00              6,00              0,00               258                 2
75%             0,00              6,00              0,00               535                 5
95%             0,00              8,00              0,00              1916                24
98%             0,00             11,72              0,00              3311                50
99%             0,00             28,46              0,00              5722                72
Min             0,00              2,00              0,00               104                 0
Max             0,00          45359,00              0,00          14530764            182785

What this basically means is that 99% percentile of our partitions are small (5KB) while the biggest is 14MB! That’s a huge difference and clearly shows that we have a hotspot on a partition somewhere.

So now we know for sure that we have an over represented key in our data set, but what key is it and why?

The culprit

So we looked at the cardinality of our data set keys which are SHA256 hashes and found out that indeed we had one with more than 1M occurrences while the second highest one was around 100K!…

Now that we had the main culprit hash, we turned to our data streaming pipeline to figure out what kind of event was generating the data associated to the given SHA256 hash… and surprise! It was a client’s quality assurance bot that was constantly browsing their own website with legitimate behaviour and identity credentials associated to it.

So we modified our pipeline to detect this bot and discard its events so that it stops polluting our databases with fake data. Then we cleaned up the million of events worth of mess and traces we stored about the bot.

The aftermath

Finally, we cleared out the data in Scylla and tried again from scratch. Needless to say that the curves got way better and are exactly what we should expect from a well balanced cluster:

Thanks a lot to the ScyllaDB team for their thorough help and high spirited support!

I’ll quote them conclude this quick blog post: