Downtime problem

[juc1]

Hi all

I use the pingdom monitoring service. The uptime on my Debian 5 VPS used to be very good but then it looks like something went wrong on September 21 because since then my two websites have had constant interruptions - downtimes of 5, 10 and 15 minutes almost every day - 27 days out of 30 in October so far - and sometimes 4 or 5 times per day. There is no reason for this such as traffic spikes, scheduled back ups etc. I believe that the pingdom alerts are generally accurate ie not false positives.

I submitted a ticket about it but the reply was very unhelpful saying only that according to the 'uptime' command the VPS was up. But of course there might be some other problem such as network connectivity etc. Anyway - can I get anyone to look at this? Should I just submit another ticket - to be honest I don't have much confidence in some of the tech support staff. Or should I just abandon this VPS and move to another one?

Thanks...

October uptime report attached

103

[juc1]

Here is a 'top' command from today which corresponds to a pingdom down alert. The average load climbs from low to over 7 and then begins to fall again - websites fail for about 15 minutes or until I restart the VPS...


104

[juc1]

Would the apache logs be any help to identify the cause of the problem?

[jeremya]

Would the apache logs be any help to identify the cause of the problem?

Yes please :)

Also, what kind of application is running? Is it making using of MySQL or another database? Is everything involved with the website hosted directly on your VPS?

[juc1]

Yes please :)

@ jeremya thanks for your reply. I am not familair with apache logs so which files would be helpful? This is my list of files in /var/log/apache2/ apache logs - Pastebin.com (http://pastebin.com/Z0yQHzQS)




Also, what kind of application is running? Is it making using of MySQL or another database?

It is a drupal website with MySQL database.


Is everything involved with the website hosted directly on your VPS?

Yes

Thanks

[juc1]

VPS down 4 times so far today for a total of one hour. :(

One of them I got the attached plesk error

107

[jeremya]

Hmmm, is Plesk down every time this occurs? If so, we may not find what we're looking for in the apache logs but if it's the load is getting particularly high leading up the mark it might help us.

We would want to see the error.log.##.gz with a timestamp just after or during a given incident date.

[jeremya]

Out of curiousity, is your VPS control panel (virtuozzo power panel) functioning through these issues?

[jeremya]

My first guess (unless you're also having trouble with Virtuozzo Power Panel at the same time) is that you may be a bit low on RAM. Please let us know how things are going now.