Ran YUM upgrade. site now down & plesk wont start. donation

[jayvin]

hi guys, thanks for the help.

I had to re-image server before you posted your suggestions. I'm now firmly stuck on plesk 9.5.2 which I cannot update via 'yum upgrade' or using the software updater admin panel within plesk. Both break my plesk panel and bring my sites down. Ive had to re-image the server twice in 3 days.

Im now not sure if I should be using yum upgrade to update plesk or not?

I also posted the problem on the plesk forums http://forum.parallels.com/showthread.php?t=106245

the last post links to other threads with people having similar problems updating plesk 9.5.2.

I ran rpm -q psa and got this: psa-9.5.2-cos5.build95100504.10

as usual the parallels forum support is next to non existent (from being plesk user for some years I've found if Igor doesn't help, you may as well give up on the forum) and as they vet every single forum post before it's shown it doesn't create a quick level of community support.

I really want to be able to update to the latest versions of yum and plesk but Im afraid if I attempt another upgrade ill spend another 3 stress filled days with my site down.

I'm no developer and learning as I go, Id love a simple guide of how and what I should be implementing on my server to keep my software updo date. i.e. how I should be using 'yum' and plesks own update software admin.

I can provide as much info as I can.

[BruceLee]

Haven't you tried my suggestion? It looks to me that the last initscript update causes your network card not to be bound/found. And therefore almost everything is not working. I don't think it's Plesk related.

If I were you I would exclude the last initscript update in repo file and than update.
But as always it's just a guess from my side.

[scott]

Can you set up a test environment? We use vmware and kvm here for that kind of thing.

[nhoeller]

I had exactly the same problem on November 21st after a yum update on CentOS 5 (Virtual Private Server). I saw nothing in the yum update list affecting Plesk directly, although it did update mysql and mysql-server. I ended up doing a full restore to get my websites back online and only captured basic log information. The messages log looked very different from other restarts:

Code: Select all

2010-11-21T11:13:17.607933-05:00 s15340421 xinetd[11299]: xinetd Version 2.3.14 started with libwrap loadavg labeled-networking options compiled in.
2010-11-21T11:13:17.608005-05:00 s15340421 xinetd[11299]: Started working: 2 available services
2010-11-21T11:13:26.717002-05:00 s15340421 saslauthd[11895]: detach_tty      : master pid is: 11895
2010-11-21T11:13:26.717121-05:00 s15340421 saslauthd[11895]: ipc_init        : listening on socket: /var/run/saslauthd/mux
2010-11-21T11:13:26.737512-05:00 s15340421 avahi-daemon[11914]: Found user 'avahi' (UID 70) and group 'avahi' (GID 70).
2010-11-21T11:13:26.737720-05:00 s15340421 avahi-daemon[11914]: Successfully dropped root privileges.
2010-11-21T11:13:26.738238-05:00 s15340421 avahi-daemon[11914]: avahi-daemon 0.6.16 starting up.
2010-11-21T11:13:26.738607-05:00 s15340421 avahi-daemon[11914]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!
2010-11-21T11:13:26.738917-05:00 s15340421 avahi-daemon[11914]: dbus_bus_get(): Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory
2010-11-21T11:13:26.738977-05:00 s15340421 avahi-daemon[11914]: WARNING: Failed to contact D-Bus daemon.
2010-11-21T11:13:32.147159-05:00 s15340421 init: no more processes left in this runlevel
2010-11-21T11:18:02.775739-05:00 s15340421 poplock DB cleanup[15458]: fatal error: postmap: fatal: could not find any active network interfaces
2010-11-21T11:28:02.320672-05:00 s15340421 poplock DB cleanup[19659]: fatal error: postmap: fatal: could not find any active network interfaces
2010-11-21T11:29:19.145406-05:00 s15340421 shutdown[19738]: shutting down for system halt
2010-11-21T11:29:19.163393-05:00 s15340421 init: Switching to runlevel: 0
2010-11-21T11:29:19.190977-05:00 s15340421 saslauthd[11895]: server_exit     : master exited: 11895
2010-11-21T11:29:27.167910-05:00 s15340421 xinetd[11299]: Exiting...

I was unable to PING the server IP, suggesting a failure in initializing networking. So far, Internet searches have not narrowed down what the problem might be.

My hosting provider said that updating mysql can cause Plesk failures. This particular problem seems to be more basic. Also, mysqld showed up in the process list and the Virtuzzo control panel claimed it was successfully started. However, psa was not started, and starting psa through Virtuzzo resulted in messages claiming the mysqld start failed. mysqld.log showed no obvious error messages.

There were a number of Plesk-related tasks that did not show up in the process list. /usr/sbin/sw-cp-serverd was also not running - not sure if that is germane to this issue.

The good news is that I was able to recover from the problem in less than 3 hours. The bad news is that I have no idea how to avoid this kind of problem in the future, nor do I know how to identify which component is causing the problem. I can manually install one component at a time but may very well end up in the same situation where I cannot access the server other than in repair mode. It looks like my best bet is Michael Shinn's suggestion to chroot /repair and see if I can do a yum downgrade on the failing component.

[scott]

If you upgrade from mysql 5.0 to 5.1 you should read this:

http://www.atomicorp.com/wiki/index.php/Mysql

[nhoeller]

It looks like I am currently running mysql/mysqld 5.0.77 (after the restore back to the point before I did the yum update). According to the log, yum update installed mysql-5.0.77-4.el5_5.4.x86_64.rpm and mysql-server-5.0.77-4.el5_5.4.x86_64.rpm.

I suspect the apparent issue with mysqld after the yum update was a red herring.
thanks, Norbert