Our www2 reseller server will be brought down for maintenace on the 1st of April, 2004 (no, not an April fools joke) for routine maintenance. Downtime is expected to be 5-10 minutes, during this 8 hour window. We will try to push this back as far as possible, we will be adding another 1/2 GB of RAM to help alleviate recent speed issues in relation to the many, many MySQL processes the server handles.
Let me know if you have any questions, as always .
Ok, well that upgrade certainly went anything but smoothly.
Heres the summary of what happened:
- www2 was gracefully shutdown, 512MB of DDR ECC Registered RAM was added
- www2 came back up, ran for about 15 minutes, then pagefaulted (crashed)
- RAM was moved to another slot, www2 came back up, ran for about 30 minutes. It was noticed that apache/php was broken from the first attempt (still uncertain as to how), specifically one of the php libraries, php was disabled as apache was recompiled
- Recompile finished, no impact, the server then crashed
- RAM was decided to be bad after 2 failures in a row, removed, and labeled as DOA. Server was rebooted without it only to find both apache and php non-functional
- Last few hours have been spent debugging this, currently apache is up, php is not. Basic apache compilations even failed, resulting in continual segfaults (aka no webserver).
This issue *will* be resolved before the sun is up, hopefully before midnight.
It would appear that the error, though it was writing nothing to the /usr/local/apache/logs/error_log was being caused by php, more specifically, the Zend Optimizer portion, or its libraries.
After installing a fresh copy of Zend Optimizer based on a hunch, apache has restarted flawlessly with php support.
For future notice (and anyone searching the forum if they get stuck down the road), the following error *was* fixed by reinstalling zend optimizer, not repairing apache/php/etc:
[Thu Apr 1 23:28:31 2004] [warn] NameVirtualHost 66.199.181.199:80 has no VirtualHosts
/usr/local/apache/bin/apachectl: line 192: 7210 Segmentation fault (core dumped) $HTTPD -DSSL
/usr/local/apache/bin/apachectl startssl: httpd could not be started
If anyone has any comments, questions, or concerns, please do *not* hesitate to email me immediatly regarding this matter, or any related matter. Anyone whose account has not been kept up to our 99.5% uptime guarantee as a result of this should immediatly contact me regarding this also. Everyone affected by this has my appologies, quick hardware swap gone bad would be a fitting title to this post.
.