Hi Matt,
Just a quick thought; can you get timing information - not an expert on
MySQL but this might do something:
http://dev.mysql.com/doc/refman/5.6/en/performance-schema.html
Or do you need to vacuum it or something?
cheers
-j
On 17/08/2015 10:34, Matt Doidge wrote:
> Heya all.
>
> I have yet to collect any actual evidence yet so this is all very
> hand-wavey, but I swear our DPM is show it's age. The hardware seems
> okay, it's on it's third year of a 5 year warranty and was generously
> specced and showing no errors. But I have a feeling that we're seeing
> an increase of timeout errors, particularly on the srm2.2 service.
>
> There was a time when I wouldn't see a Ops srm test timeout failure
> from one month to the next, now I swear I'm seeing them weekly. Same
> with job failures where there was a transfer error but I can't see
> anything wrong. We're not significantly busier now then we were 2
> months ago. Of course I could be suffering from observation bias.
>
> I have a day and a bit's downtime in a fortnight, so there's the
> opportunity for some soothing to be done. But what? My first thought
> is doing something about the database:
>
> # du -sh /var/lib/mysql/
> 31G /var/lib/mysql/
>
> Perhaps a simple databse dump and reinitialise from the aforementioned
> dump. Anyone else have any other suggestions? Or perhaps other
> performance-detracting things to look out for? We could put them in
> the wiki or a blog post :-)
>
> Thanks in advance,
> Matt
|