While we are at it, rename disk-checkup.nix to btrfs-auto-scrub.nix and
move it into //ops/modules. I originally wanted to have additionally
disk health related services in that module, but the btrfs scrub
functionality is nicely self-contained and reusable, so I think it makes
sense to have this in a more central location.
Change-Id: Iabdd62838eef009540ca71abafd921afda2a9b47
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10128
Reviewed-by: sterni <sternenseemann@systemli.org>
Autosubmit: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Network configuration and initrd setup is basically the same as with
edwin, but we are using md for Software RAID this time as well as LVM
over two partitions with LUKS:
- sda2 <-- RAID1 --> sdb2 (boot-raid)
└ boot partition, ext4 (encrypted-container-raid)
- sda3 <-- RAID1 --> sdb3
└ LUKS container
└ Volume Group vgmain
├ Logical Volume vgmain/swap
│ └ swap
└ Logical Volume vgmain/root
└ btrfs
So we no longer rely on btrfs raid1 due to question marks over its
reliability (I personally did not have any problems though). This also
means that we have less LUKS containers we need to unlock when
booting (kind of neglible improvement). The biggest improvement is that
we have redundancy for the swap, so a disk failure shouldn't cause
memory corruption/loss.
Change-Id: I14f065b659857415917d9a60a7ec019e687f8d1c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10127
Tested-by: BuildkiteCI
Autosubmit: sterni <sternenseemann@systemli.org>
Reviewed-by: tazjin <tazjin@tvl.su>
Reviewed-by: sterni <sternenseemann@systemli.org>
These settings would also be applied to a machine that'd replace edwin,
so it's useful to have them outside edwin's default.nix.
Change-Id: I4e8f464118a103645e53909a87c6ee4446022fa3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10125
Reviewed-by: sterni <sternenseemann@systemli.org>
Autosubmit: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
This introduces a new structure, as activities are moved to a separate
AWS Account for reasons (tm).
Change-Id: Ic310eca2dc0d4ee81bae8944238b26910067336a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10124
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This removes everything related to workspaces, as well as the current
RandR configuration, creating a (cleaner) slate for switching to tabs.
This was supposed to be committed earlier, but got broken by
clbot (see b/333).
Change-Id: I2d110bca0d6629a505699210f0aba12882f83d48
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10115
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Sets up tab-bar-mode for EXWM, and adds simple initial configuration
and shortcuts to make it behave sanely. I had to assign tab switching
shortcuts myself, as the EXWM input key stuff doesn't work correctly
with the built-in way of setting these shortcuts in tab-bar-mode.
Change-Id: Icd96f03ae7bd2bd5f7c2d59f9aca7a1ca2e7f788
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10116
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Adds an `exwm-assign-workspaces` function that automatically creates
and assigns workspaces to each currently connected monitor. The first
workspace (index 0) is always on the primary monitor.
This function should be idempotent and can be called at any point to
synchronise X outputs and what EXWM is displaying on them.
This works because tabs are disconnected from workspaces completely,
so I don't have to care about what's going on on other workspaces
anymore.
Still missing:
* functions to connect/disconnect outputs
* switching to other outputs from within emacs commands (i.e. without
the mouse)
Change-Id: I7c24aa1b45218fe35de6939e799852b5d11d1272
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10119
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
This one is a bit stupid because I couldn't figure out a way to
determine the active workspace. It's definitely possible (either
through some XCB calls, or through state management in screen change
hooks), but for now this is fine.
Change-Id: I5e4c531b248caa0021664bad9dc196bef60cfbac
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10122
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
By default, this will be the (only) remaining monitor. In N>2
situations, ask the user.
Change-Id: Id68fcf60d56d0414d1072b8ffeef72c608678f52
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10121
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Uses a bunch of weird xrandr invocations and completing reads to
configure screens the way I want.
Note that this has a known bug where disconnecting a primary screen
will *not* make one of the remaining screens primary.
Change-Id: Ide5322df446685cc4740d4ddd7b6ca8682375050
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10120
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Something changed in the machinery and broke the overriding. I didn't
notice this, as a I was temporarily using an unpatched EXWM.
Change-Id: I1a4e8ea63bd116d86a430e680c2b631474e9a0fe
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10047
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Make these keys behave a bit more like the Fn+F* keys on the thinkpad
keyboards.
Seems there's currently no trivial way to get mic mute, keyboard mute
and wifi toggle sent out, but considering the thinkpad usb keyboard is
able to, this should be possible somehow here too - but not today, left
for a followup.
Change-Id: I529a958c78116dd9f7250c938e2e7989b296d6c6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10076
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
It seems the regex is not perfect, it choked on a single log line:
```
Nov 13 03:10:19 archeology-ec2 59nkrwmih3ywaxrgxqj79pn395fs6m17-parse-bucket-logs-continuously[11105]: Code: 117. DB::Exception: Line "d57bd890fbd1ae16625bdb8168064125e013198099b7e1b3c24878a4d03c3ab8 nix-cache [12/Nov/2023:09:13:02 +0000] xxx.xx.xxx.xxx - VB7SJVZ108DSSN67 REST.POST.OBJECT index.html "POST /index.html HTTP/1.1" 405 MethodNotAllowed 348 - 4 - "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36" - 0bFdGKbi0n9JHXU1a2hijcJwmYdc6lG2xgbdozc3wS6mlUkBE7ssrQCHIDdOLebo78o2cGbhivY= - ECDHE-RSA-AES128-GCM-SHA256 - nix-cache.s3.amazonaws.com TLSv1.2 - -" doesn't match the regexp.: (in file/uri log/2023-11-12-10-19-50-80805A702ECF65EB): (at row 5)
```
This was due to the user-agent field. The regex is now fixed.
The request itself is fun (someone trying to POST an index.html to the
bucket), and we should probably filter this on the Fastly side already,
not via IAM,
In any case, there's no point failing to parse if a single line doesn't
match the regex - we can just skip them.
For the sake of completeness, logs for that day have been reprocessed
and reuploaded.
Change-Id: Id98a7167a381cda06d150ad5118ee9e70ead277e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10034
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
I no longer work at ReadySet.
Change-Id: Idc19e2d68846551b6cd94f84594712692ebe35a9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9976
Tested-by: BuildkiteCI
Autosubmit: grfn <grfn@gws.fyi>
Reviewed-by: grfn <grfn@gws.fyi>
This packages up my keyboard firmware used for the Keychron K6 Pro.
We add a custom keymap to the `keyboards/keychron/k6_pro/ansi/rgb/
keymaps` directory, a copy from the `default` one (with a modified
`keymap.c`), and then build that as a makefile target.
`via` is *disabled*, as their keybindings take priority over keymap.c.
Luckily, only `qmk` seems to be sufficient to build it.
A simple `:flash` target/script is provided as well, it relies on some
udev rules set in the global system
(`hardware.keyboard.qmk.enable = true`).
Change-Id: I9f7a7a992e13516c32033127f94e37aec62d6b67
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10020
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
This allows using awscli inside a shell.
Clickhouse AWS SSO integration still seems broken unfortunately, even
with https://github.com/ClickHouse/ClickHouse/pull/54347 included in
our bump - it seems it's coming up with another token file path than the
AWS SDK:
> SSOCredentialsProvider: Unable to open token file on path: /home/flokli/.aws/sso/cache/da39a3ee5e6b4b0d3255bfef95601890afd80709.json
This is the sha1sum of the sso_start_url, not the sha1sum of the
session-name (nixos / f2f059b8b7298f1ad52636d67cef8b719aa83bf5).
Change-Id: Ia1bdec03c4f269a7415c42c90c1f4fd3d928f770
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10012
Reviewed-by: edef <edef@edef.eu>
Tested-by: BuildkiteCI
This was previously only in my Telegram channel, but it might as well
be on the blog itself.
Change-Id: I301ebeaa4dd1875f3858cee5259a5c689b950790
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10009
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
This adds a `parse-bucket-logs.{service,timer}`, running once every
night at 3AM UTC, figuring out the last time it was run and parsing
bucket logs for all previous days.
It invokes the `archeology-parse-bucket-logs` script to produce
a .parquet file with the bucket logs in `s3://nix-cache-log/log/` for
that day (inside a temporary directory), then on success uploads the
produced parquet file to
`s3://nix-archeologist/nix-cache-bucket-logs/yyyy-mm-dd.parquet`.
Change-Id: Ia75ca8c43f8074fbaa34537ffdba68350c504e52
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10011
Reviewed-by: edef <edef@edef.eu>
Tested-by: BuildkiteCI
Clickhouse also has column compression, configurable with the
output_format_parquet_compression_method setting.
It defaults to lz4, and the previous setting got a a zstd-compressed
parquet file with lz4 data.
Set output_format_parquet_compression_method to zstd instead, and sort
by timestamp before assembling the parquet file.
The existing files were updated to the same format with the following query:
```
SELECT * FROM file('bucket_logs_2023-11-11*.pq', 'Parquet', 'auto') ORDER BY timestamp ASC INTO OUTFILE 'bucket_logs_2023-11-11.parquet' SETTINGS output_format_parquet_compression_method = 'zstd'
```
Change-Id: Id63b14c82e7bf4b9907a500528b569a51e277751
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10008
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Tested-by: BuildkiteCI
This adds a `archeology-parse-bucket-logs` CLI tool to `$PATH`.
It can be invoked like this:
```
archeology-parse-bucket-logs http://nix-cache-log.s3.amazonaws.com/log/2023-11-10-00-* bucket_logs_2023-11-10-00.pq.zstd
````
… and will produce a zstd-compressed Parquet file for (roughly) that
time range.
As the EC2 instance credentials don't give access to the logs bucket
(yet), other AWS credentials need to be provided.
This can be accomplished by using "AWS_ACCESS_KEY_ID",
"AWS_SECRET_ACCESS_KEY", "AWS_SESSION_TOKEN" from
"Option 2: Manually add a profile to your AWS credentials file (Short-
term credentials)" in AWS IAM Identity Center.
Processing logs for a one-hour range takes a minute or two, the
resulting zstd-compressed Parquet file is around 40-80M in size.
Processing logs for a whole day takes some 25mins, due to the sheer
amount of data (12 GB of raw log data, distributed among 450k individual
files, 20Mio log lines), but at least clickhouse isn't able to parse the
resulting parquet file back in:
> Code: 36. DB::Exception: IOError: Couldn't deserialize thrift: MaxMessageSize reached
For future automation tasks, it's probably better to run this once an
hour, and further join the data later on.
Change-Id: I6c8108c0ec17dc8d4e2dbe923175553325210a5c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10007
Tested-by: BuildkiteCI
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Rather than picking up from clickhouse-specific config files, this gets
it to pick up from the ambient environment, which is closer to (but not
the same as) the AWS default credentials chain.
Change-Id: I9c498c231974ed345c3e3d354ec230052b4d0ff2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10006
Tested-by: BuildkiteCI
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
This behaviour might change (or not), see https://github.com/ClickHouse/
ClickHouse/pull/42003, but as of now, a `--progress` will provide some
progress.
Change-Id: I4891b6e2f96f2656858e71f88a226d24f0d45dc3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10005
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Tested-by: BuildkiteCI
First pass at an xdotool-based command to edit the current text input in
emacs
Change-Id: I1e04612478292fe83083d197d481e034a9fce97f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9971
Reviewed-by: grfn <grfn@gws.fyi>
Autosubmit: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
This can apparently work around some of the CPU throttling bugs on
~modern~ computers.
Change-Id: I807ece85d3eba53857a1cb1e73a33f7924538e96
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9895
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Expose `deps` separately, add a direnv with PATH_add for it to bring
tooling into $PATH.
Change-Id: I432cd2b082cad89e08bef78dc4653e10e137cd6b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9842
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Avoid having to re-enter the shell whenever the config is changed.
Change-Id: Ib9f6bb4075e29acaeb4863d64c017695ca85b60b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9841
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
