[Solved] Web server keeps dying after upgrade to RC4

[robmorin]

So after a very long weekend, of fixing this and that after my upgrade from rc3 to rc4, i think a few things should be added to the upgrade procedure faq or guideline.
I also think that a warning should be issued that when running the apt-get upgrade for Etch that apache prefork is removed and worker is installed thus removing libapache2-PHP4, thus breaking some web apps

i think just before the step that says to copy over /usr & /var there should be a note in there to remind folks that everything they might have customized like, zone files, webmail stuff, apache config stuff, and and anything in /etc/ispcp/* will be wiped out and notes and or backups should betaken, its never a bad idea to tell peole to backup many times in a step by step guide...

Before running ispcp-update a note should be given stating how long it will take to do and once again warning any custom changes will be wiped out. I have a bit over 1000 mailboxes and about 20 domains and it took a little bit over an hour for the update script to finish...

These are not complaints but just suggestions i still love Omega! but i would not someone thinking upgrading is as easy as some folks seem it to be be. I guess if your running STOCK Omega and never make custom changes, your ok.

Last but no way least is it did break my Apache and i cannot find the issue, before i post a plethora of logs, i thought it best to ask which logs best to post, suexc.log and apache maybe?

My problem is apache sort of hangs every 60 mins or so. a ps shows apache still running but it does not answer any requests, a simple restart fixes everything..... so i did a cron to restart each 15 mins during the last weekend as i needed a break..

I figure it must be a fastcgi thing since prefork was replace with worker and fastcgi_ispcp.conf was replaced too...

Any ideas or what else should i look for...

Thanks...

robmorin Wrote:i found an issue with mysql restarted it and almost all is working now...

However i have a few sites that do not work, mostly drupal sites.. i see this in the error.log

[Fri May 30 12:55:18 2008] [debug] proxy_util.c(1589): proxy: grabbed scoreboard slot 141 in child 23610 for worker http://localhost/stats/domain.com
[Fri May 30 12:55:18 2008] [debug] proxy_util.c(1608): proxy: worker http://localhost/stats/domain.com already initialized
[Fri May 30 12:55:18 2008] [debug] proxy_util.c(1672): proxy: initialized worker 141 in child 23610 for (localhost) min=0 max=25 smax=25
[Fri May 30 12:55:25 2008] [error] [client 69.196.5.28] FastCGI: comm with (dynamic) server "/var/www/fcgi/domain.com/php4-fcgi-starter" aborted: (first read) idle timeout (300 sec), referer: http://www.domain.com/
[Fri May 30 12:55:25 2008] [error] [client 69.196.5.28] FastCGI: incomplete headers (0 bytes) received from server "/var/www/fcgi/domain.com/php4-fcgi-starter", referer: http://www.domain.com/
[Fri May 30 12:57:13 2008] [error] [client 69.196.6.212] FastCGI: comm with (dynamic) server "/var/www/fcgi/domain.com/php4-fcgi-starter" aborted: (first read) idle timeout (300 sec), referer: http://www.domain.com/
[Fri May 30 12:57:13 2008] [error] [client 69.196.6.212] FastCGI: incomplete headers (0 bytes) received from server "/var/www/fcgi/domain.com/php4-fcgi-starter", referer: http://www.domain.com/

Any help appreciated... keep in mind thsi was after an RC3 to RC4 upgrade done last night...

Thanks
Rob..
i need to go out for a bit i will be back in a bit

Zothos Wrote:sure. But without logs/errors or an ssh account nobody can help you...