[Engineering] Analytics hosts reboot announcement - Wed 17th 10 AM CET

Luca Toscano ltoscano at wikimedia.org
Mon Jan 15 08:50:28 UTC 2018


Hi everybody,

as you already know we are deploying the new Linux kernel to fix the
Meltdown vulnerability across the production fleet. This means that I need
to reboot all the stat boxes (stat100[456]) and also analytics1003 (running
Oozie, Camus, Hive, etc..), probably interfering with the work that you are
doing on tmux/screen sessions :)

The maintenance is planned for Wed 17th at around 10 AM CET.

These will be the side effects:

1) All the tmux/screen sessions on stat100[456] will be killed by the
reboots, so please reach out to me or the Analytics team if you have
important work that cannot be stopped/resumed, we'll try to do our best to
reschedule the maintenance accordingly.
2) A lot of Hadoop cluster activities (like regular Oozie jobs, Hive
queries, etc..) will need to be stopped to allow a smooth reboot of
analytics1003. This might cause some of your queries to fail, again please
reach out to me if this is a problem for your work.

Thanks a lot and sorry for the trouble!

Luca (on behalf of the Analytics team)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.wikimedia.org/pipermail/engineering/attachments/20180115/1e74872b/attachment.html>


More information about the Engineering mailing list