apache hangs all the time - sugarcrm downtime

[zmrdi]

hi all

we've got sugarcrm 5.2.0i running on windows server 2003 using apache 2.2.17 and mySQL 5.1.50 on another windows 2008 server. further we use mysqldump to backup up the mysql database every 24 hrs.

now we're faced with some stability issues, seems to be caused by apache and the mysqldump.

what happens? every day at almost the same time (varies within 1,5 hrs) apache hangs and does not come back. from this time it is no longer possible to login to sugarcrm, even if apache service is running and port 80 is opened. when it happens in most cases the following message is generated in the apache log

xx.xx.xx.xx - - [29/Nov/2010:21:50:11 +0100] "GET /index.php?module=Home&action=index HTTP/1.0" 503 265

first we thought that the dump is not the trigger, but we now disabled the mysqldump and the problem did not appear. that's actually weird because the issue also happened 10 to 30 minutes BEFORE the dump starts automatically.

AFAIK http error 503 means that the service (in this case the databsase service) is not available.

honestly I have no idea how to proceed now... does someone else have a brilliant idea what to check?

I can not figure out that it is a performance issue... the mysql server is a dedicated brand new server with 2,4 GHz Xeon CPUs (8 cores all up), 18 GB of RAM (64 bit OS of course). mysql data is stored on a brand new SAN storage with gigs of cache and thousands of IOPs ... so IMHO the performance of the server is not the bottleneck in that case. We also tried to install apache on the same server that has mySQL installed, it's the same bahaviour

I really appreciate any hint or ideas...