Plesk Takes Down Server After Upgrade

[PGT92]

I have a dedicated server with Plesk pre-installed. I attempted to update Plesk from 8.0 to 8.2.1 on Saturday. I received an email notice letting me know that it failed. At that point, email was down and I wasn't able to log into Plesk (HTTP 500 Error).

Monday morning, my host company fixes the problem and finishes the upgrade 'successfully', although email is working now, other problems have occurred.

Bad:
-Plesk is 'Not found' (no more HTTP 500 error)
-All websites are down

Good:
-Email works
-FTP works
-SSH works

My host told me Monday that ports 80 and 8443 are not listening properly. Tuesday (today), upon further study, they've told me that the Plesk upgrade 'bombed' and screwed up the httpd.conf file. They've told me that they want to reinstall Plesk and see if things work then. That isn't a problem, however they told me that I would need to backup all my data since the reinstallation may wipe out some important data. I don't have the time to set all my websites back up, so anything that may cause a problem like that is out of the question.

My question is, what could be causing this problem and is reinstallation the only option?

Thanks for any help.

 

[PGT92]

I received an update from my host. They're saying the problem is much deeper and will require a complete installation of the operating system as well, and are listing hundreds of dollars worth of fees.

Does this sound right at all? Aside from a loss of connectivity to port 80 and 8443, everything else seems to be running fine.

 

[dirty@]

it is very difficult to say anything, just because the problem reason is unknown and you didn't provide any details..

i think os reinstallation that you were suggested may only be needed if there are a lot of broken system libraries and binaries.. for example, if you mistakenly upgraded system packages with updates from absolutely other operating system..

if the problem is in apache only and was caused by corrupted apache files only, then no os reinstallation is needed of course..

unfortunately there are no more-less detailed information from you to suppose what may be done in your case.

 

[PGT92]

I upgraded Plesk via the console's updater application and let it install what it suggested. I made no changes. Originally, the Qmail portion of the update failed due to a 'permission problem'. My host solved the problem, and finished the update on their own, which was 'successful'. Email came back up, but the websites went down.

It just seems to me that my host's techs are being lazy in their diagnosis and are trying to go the easy route with a complete reinstallation... which would be very expensive and time consuming on my end.

I'm not very familiar with Linux, so I'm not sure how much more detailed I need to be.

Thanks.

 

[Amin Taheri]

If I remember correctly

/usr/local/psa/admin/bin/websrvmng -a --reconfigure-all

rebuilds the httpd.conf file that psa uses and rebuilds all client httpd.include files for each domain.

That should solve the problem of any of those two things being broken or incorrect/incomplete.

 

[PGT92]

Originally posted by HostingGuy
If I remember correctly

/usr/local/psa/admin/bin/websrvmng -a --reconfigure-all

rebuilds the httpd.conf file that psa uses and rebuilds all client httpd.include files for each domain.

That should solve the problem of any of those two things being broken or incorrect/incomplete.

I tried and nothing happened.

 

[PGT92]

From what I understood, Apache was running the whole time, however, tonight I realized it wasn't. When attempting to start Apache, I received the message:

[Wed Oct 10 21:04:48 2007] [warn] NameVirtualHost <snip>:80 has no VirtualHosts
[Wed Oct 10 21:04:48 2007] [warn] NameVirtualHost <snip>:443 has no VirtualHosts
[Wed Oct 10 21:04:48 2007] [warn] NameVirtualHost <snip>:80 has no VirtualHosts
[Wed Oct 10 21:04:48 2007] [warn] NameVirtualHost <snip>:443 has no VirtualHosts
(98)Address already in use: make_sock: could not bind to address <snip>:443
no listening sockets available, shutting down
Unable to open logs!


Side question... is it even possible for a Plesk update to do so much damage to a server as to require a complete OS reinstallation?

 

[PGT92]

One more thing...

Inputting the command 'apachectl stop' yields:

[Wed Oct 10 21:17:48 2007] [warn] NameVirtualHost 172.20.80.226:80 has no VirtualHosts
[Wed Oct 10 21:17:48 2007] [warn] NameVirtualHost 172.20.80.226:443 has no VirtualHosts
[Wed Oct 10 21:17:48 2007] [warn] NameVirtualHost 64.41.72.226:80 has no VirtualHosts
[Wed Oct 10 21:17:48 2007] [warn] NameVirtualHost 64.41.72.226:443 has no VirtualHosts
httpd (no pid file) not running

However, 'service httpd status' yeilds:

httpd (pid 26608) is running...

I restarted httpd and although I got the four 'NameVirtualHost' warnings, it restarted.

I also noticed an odd change after I commented the line 'Listen <snip>:80' out of my httpd.conf file, although I'm not sure how much it matters... previously, any attempt to go to one of my websites cause the browser to simply hang for several minutes. Now, going to one of my websites pulls up a 'not found' page immediately. Does this mean anything?