[Labs-l] wikidata-dev-3 instance causing 80M/s of disks writes

Petr Bena benapetr at gmail.com
Sat Jan 5 10:16:43 UTC 2013


howcome sed is writing to disk? I suppose it was a part of other command?


On Sat, Jan 5, 2013 at 3:29 AM, Ryan Lane <rlane32 at gmail.com> wrote:

> Thanks!
>
>
> On Fri, Jan 4, 2013 at 6:02 PM, aude <aude.wiki at gmail.com> wrote:
>
>> On Sat, Jan 5, 2013 at 2:42 AM, aude <aude.wiki at gmail.com> wrote:
>>
>>> On Sat, Jan 5, 2013 at 12:52 AM, Ryan Lane <rlane at wikimedia.org> wrote:
>>>
>>>> There's a weird sed statement that runs every hour on wikidata-dev-3:
>>>>
>>>>
>>> I can't ssh into to that instance to look at this.  No problem ssh into
>>> our other wikidata-dev labs instances.
>>>
>>> I'm not exactly sure the reason we are doing this, so probably needs to
>>> wait until someone else on our team looks into it.
>>>
>>
>> Alright, I was able to ssh into the instance.  We have disabled the
>> cronjob for now, and checking with our team if we need still need this or
>> not or what.
>>
>>
>>>
>>> Cheers,
>>> Katie
>>>
>>
>> [snip]
>>
>>
>>> This causes performance issues for other virtual machines in the same
>>>> host. Can the wikidata folks please handle this some less disk intensive
>>>> way?
>>>>
>>>> - Ryan
>>>>
>>>> _______________________________________________
>>>> Labs-l mailing list
>>>> Labs-l at lists.wikimedia.org
>>>> https://lists.wikimedia.org/mailman/listinfo/labs-l
>>>>
>>>>
>>>
>>>
>>> --
>>> @wikimediadc / @wikidata
>>>
>>
>>
>>
>> --
>> @wikimediadc / @wikidata
>>
>> _______________________________________________
>> 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: <http://lists.wikimedia.org/pipermail/labs-l/attachments/20130105/6d0a8b12/attachment-0001.html>


More information about the Labs-l mailing list