tl;dr: we are about to delete a whole lot of files on PAWS. Please reach
out /quickly/ if you have unusually large storage needs in paws or if
you find important things missing.
--
Due to an increase in inappropriate and/or abusive uses, we are setting
up more aggressive scanning and clearing of the NFS storage that
underlies PAWS notebooks. This will be largely automated and will be a
fairly blunt instrument; ideally we will only erase files associated
with invalid uses but there are likely to be some false positives so I'm
hoping to get user feedback when we overstep.
In order to free up working space I'm going to need to delete some large
files right now. Going forward, files will be moved into purgatory for a
couple of weeks before deletion, so if you find important data missing
it can likely be recovered if you contact us right away. Either look for
an admin on IRC or follow up on this phab task:
https://phabricator.wikimedia.org/T327936
This is all briefly documented at:
https://wikitech.wikimedia.org/wiki/PAWS/Getting_started_with_PAWS#Usage_an…
_______________________________________________
Cloud-announce mailing list -- cloud-announce(a)lists.wikimedia.org
List information: https://lists.wikimedia.org/postorius/lists/cloud-announce.lists.wikimedia.…
One of the cloud-vps compute servers, cloudvirt1071, crashed a few
minutes ago. I have now restarted it and moved all VMs to other hosts,
but a small number of VMs suffered 15 or so minutes of downtime followed
by a reboot. Those VMs are:
kube-2js6t-default-worker-5ff26-rg5n9-nxtmp.zuul
kube-2js6t-default-worker-5ff26-rg5n9-r722f.zuul
k3s.catalyst-dev
mediawiki2latex.collection-alt-renderer
toolsbeta-test-k8s-worker-nfs-8.toolsbeta
tools-k8s-worker-nfs-44.tools
tools-k8s-worker-nfs-32.tools
No data was lost, and no action should be needed; this is just
informational for anyone who noticed the downtime. Toolforge tools
should have been unaffected, although some pods were likely rescheduled
after a few seconds of hesitation.
Any followup on this incident will be tracked on
https://phabricator.wikimedia.org/T431374
-Andrew
_______________________________________________
Cloud-announce mailing list -- cloud-announce(a)lists.wikimedia.org
List information: https://lists.wikimedia.org/postorius/lists/cloud-announce.lists.wikimedia.…
Hello,
I am facing a problem with a missing directory called "cpufreq" on the
instance mediawiki2latex on the project collection-alt-renderer
the directory:
/sys/devices/system/cpu/cpu0/
exists, but the directory
/sys/devices/system/cpu/cpu0/cpufreq
does not exist. It happened when I ran the following command:
dhun@mediawiki2latex:~$ chromium --no-sandbox --no-pdf-header-footer
--headless --print-to-pdf="map1.pdf" map1.html
[0706/114648.182905:ERROR:third_party/crashpad/crashpad/util/file/file_io_posix.cc:145]
open /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq: No such file
or directory (2)
[0706/114648.182973:ERROR:third_party/crashpad/crashpad/util/file/file_io_posix.cc:145]
open /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq: No such file
or directory (2)
Trace/breakpoint trap
So I think chromium wants to access the clock rate of the cpu but fails
in doing so. I found a post on a forum that suggests to change the
settings of the hypervisor:
https://unix.stackexchange.com/questions/676022/chrome-crashes-with-subproc…
Any help is appreciated!
Yours Dirk