The daily 1pm BST site crash...

**talkpoliticsuk66** · 05-09-2011 07:49 AM

This has been reported by another user from the UK, last week, but...

Every single fragging day, my site - which is on Boru - become inaccessible from around 1pm local time, which I think is about 7am at the server location if the geolocation for Boru is correct in placing it physically in the Chicago area.

Every single day, the same issues - either 'page temporarily unavailable' or a failure to connect to MySQL, lastiing from one to two hours before it clears up of its own accord.

To me, as an ex-sys admin that says system overload - something is hammering the MySQL server at the same time every single day, which is no biggie for users on your side of pond, but peak lunchtime viewing hours over here.

Obvious possibilities would be something internal like a main server backup but the timing also raises the possibility that there's just a shedload of sites on the server all running Cron jobs at the same time, most likely database backup scripts, and this is just chewing up all the connections.

Can we at least get this investigated to find out what's hammering the MySQL server so hard at the same time every day - it shouldn't be difficult to pin down the cause(s) from the server logs.

For the 22 hours of the day, the hosting and reponse times are top notch for a free service, even allowing for all recent outages caused by problems with the SAN migration, its just a shame that us UK and European users are having to put with a regular daily outage during our peak hours.

**talkpoliticsuk66** · 05-10-2011 07:21 AM

Okay, its now 1:20pm BST and guess what?

No database connection again, all on schedule as it has been for at least the last couple of weeks.

**talkpoliticsuk66** · 05-11-2011 08:35 AM

And again...

Kept a watch on this today.

Started to get a major slow down in response at around 3-4 minutes past 1pm BST and by 13:16 pm I started to get cannot connect to database errors.

Some sort of script (or series of scripts) is starting up at 1pm BST and absolutely hammering the MySQL server on Boru for around 1.5 to 2 hours, making it impossible to access sites from the UK.

What I'd like to know, for starters, is what is actually causing this?

Is it an internal daily server backup or it is down to too many users scheduling their own site backups for the same time in the morning?

Without knowing the cause, its difficult to know quite what to do in order to work around the problem or even whether you're aware of this issue and are trying to sort out a fix for it, if such a fix is even possible.

I'm not the only one having this problem, so some sort of feedback would be appreciated.

---------- Post added at 01:30 PM ---------- Previous post was at 12:28 PM ----------

Okay. so we're now a full hour into today's 'scheduled' MySQL fault and I thought I'd log into my control panel to see if I could access the database through phpmyadmin.

I can't - in fact I can't even get to my control panel as all I'm getting is "Fatal error: Class 'Memcache' not found in...".

The question is whether this is the cause of the daily MySQL problem, a symptom of the problem or something unrelated.

---------- Post added at 01:35 PM ---------- Previous post was at 01:30 PM ----------

And we're finally back with access to my site, chronically slow (20 seconds+ for front end, 30 seconds+ for backend) but at least I'm getting something, even if the main account login is still screwed and showing a memcache error.