[Labs-l] ToolLabs back up

Fri Jan 30 21:20:48 UTC 2015

This wasnt a case of improvements. This was an emergency patching of a
security hole. All WMF servers had to be updated and rebooted, the only
noticeable reboots where in labs where there isnt redundancy/standby
services that can be swapped into place to prevent down time.

This is not to say that our process went without issues. I will be drafting
a postmortem from the user's perspective with several ideas and suggestions
this weekend.

On Fri, Jan 30, 2015 at 4:11 PM, Dan Andreescu <dandreescu at wikimedia.org>
wrote:

> FYI: some of us in Analytics are talking about a "graduate labs" kind of
> environment.  Which would be more stable and treated a bit less
> experimentally.  It's not something we're going to push for yet, but I was
> wondering how people here felt about it.  It feels like our current labs
> environment is improving at an amazing rate, but sometimes we might want to
> prioritize stability over improvement.
>
> On Fri, Jan 30, 2015 at 4:05 PM, Harry Burt <jarry1250 at gmail.com> wrote:
>
>> Okay, gonna do something about my tools, they fall over every time this
>> happens.
>>
>> Is it just a case of echo "webservice" > ".bigbrotherrc" ? Or does the
>> daemon have to be enabled?
>>
>> Harry
>>
>> --
>> Harry Burt (User:Jarry1250)
>>
>>
>> _______________________________________________
>> Labs-l mailing list
>> Labs-l at lists.wikimedia.org
>> https://lists.wikimedia.org/mailman/listinfo/labs-l
>>
>>
>
> _______________________________________________
> Labs-l mailing list
> Labs-l at lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/labs-l
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.wikimedia.org/pipermail/labs-l/attachments/20150130/7a79a427/attachment.html>