• The Horde webmail has been deprecated. Its complete removal is scheduled for April 2025. For details and recommended actions, see the Feature and Deprecation Plan.
  • We’re working on enhancing the Monitoring feature in Plesk, and we could really use your expertise! If you’re open to sharing your experiences with server and website monitoring or providing feedback, we’d love to have a one-hour online meeting with you.

10.4.4 autoupdate fails and deletes almost all psa files

jimbob123

New Pleskian
PRODUCT, VERSION, MICROUPDATE, OPERATING SYSTEM, ARCHITECTURE
Plesk 10.4.4, Ubuntu 10.04.3 LTS 64-bit

PROBLEM DESCRIPTION
As far as I know I had the latest 10.4.4 Plesk version and then there was a minor auto-update at 6am BST on 15/07/2012 which was supposed to keep Plesk at 10.4.4 but I assume updated some microupdate version.

This is connected to a separate thread I created: [thread]261192[/thread]

The autoupdate fails with the following error message at the end of /tmp/plesk_10.4.4_installation.log (attached):

Code:
*****  problem report *****
ERROR while trying to chown root:psaadm /opt/psa/admin/sbin/wrapper
Check the error reason(see log file: /tmp/plesk_10.4.4_installation.log), fix and try again
STOP Bootstrapper 10.13.4 post-install for BASE AT Tue Jul 17 07:07:07 BST 2012
To complete product installation one should run bootstrapper script:
      /opt/psa/bootstrapper/pp10.13.4-bootstrapper/bootstrapper.sh repair

At this point the sites on the server are still working but the following problems exist:

  • /usr/local/psa/ directory doesn't exist
  • /opt/psa/admin/sbin/ is missing almost all the files e.g. the wrapper file mentioned in the ERROR above
  • About 90% of all other Plesk files have been deleted

I copied the wrapper file across from another Plesk installation, carefully set the permissions and ownership correctly and tried to proceed as per the last line
Code:
/opt/psa/bootstrapper/pp10.13.4-bootstrapper/bootstrapper.sh repair
.

This then gave an error of:

Code:
ERROR while trying to chown root:psaadm /opt/psa/admin/sbin/mod_wrapper

So I did the same as with the wrapper file and copied the file from an existing installation and re-ran
Code:
/opt/psa/bootstrapper/pp10.13.4-bootstrapper/bootstrapper.sh repair
.

This gave an further error of a missing run-**somthing**.tar file - sorry I don't have the log files for these error messages any more.

I put the tar file in place (re-ran the bootstrapper.sh repair command) and then hit a futher error. Then I noticed that now it wasn't just Plesk Admin that had problems the entire server was knackered and the apache server was back to pointing at /var/www/ instead of the individual Plesk vhost directories.

The error message I'd got was:

"ERROR while trying to SMB database not created"

We had changed the admin password recently and it had a '-' in it - and there is a KB article suggesting this error was caused by symbols in the password. N.B. This was definitely NOT the route cause as we had the same problem when the server was restored (with an old simpler Plesk admin password) and the auto-update knackered Plesk again.

So now I tried to update the admin password but of course the script to run it in /opt/psa/admin/plib/api-cli/ was missing.

At this point I gave up and went down the route of restoring the whole server from a backup - cue an entire night spent carefully restoring files.

STEPS TO REPRODUCE

Have a 10.4.4 Plesk installation with autoupdates turned on.

After restoring a the server to a month ago the autoupdate occurred again at 6am the next morning and repeated the same error.

ACTUAL RESULT

Knackered system

EXPECTED RESULT

Non-knackered system

ANY ADDITIONAL INFORMATION

We had recently changed the Plesk Admin password - but this problem occurred twice (after restoring the whole system) with a password that had been there for quite a while and hadn't caused any problem with prior updates

Subversion is installed on the server and there are some files in the vhost directories that are controlled by subversion (apparently this can cause some conflict)

We had recently put in quite a few firewall rules restricting access to only a few IP addresses for Plesk admin, FTP and SSH. Pretty much everything except Ping was set to deny.

We have about 2 other servers with Plesk 10 on and they haven't had problems *YET*.
 

Attachments

  • autoinstaller3.log.txt
    84.7 KB · Views: 0
  • plesk_10.4.4_installation.log.txt
    10.6 KB · Views: 0
  • plesk_10.4.4_installation_problems.log.txt
    365 bytes · Views: 1
Last edited:
Hi PettyP,

Firstly - don't try to fix the admin, that only made things worse. If you've got the sites on your server still displaying then that seems to be the best you can hope for via the upgrade path.

I've had no responses at all, its very disappointing from the Parallels support. I've done my best to give as detailed description to the problem as possible - filling in the template that IgorG suggests.

Perhaps you could post the output from your /tmp/plesk_10.4.4_installation.log as well the details of your system for the Parallels support, perhaps they'll eventually do something about it.

As a company with quite a few Plesk licences we're getting increasingly disappointed by Plesk, despite them doing a pretty excellent job up until Plesk 9. Plesk 10 has been a pain from multiple perspectives.

Further to that, I can only suggest that you find some way of skipping over this update - so try and re-install Plesk with the latest version. Then turn-off auto-updates - you should get a mail from them if there are updates and then at least you can do a full server backup before installing their upgrades.

We've just left the Plesk admin in its knackered state (as the sites are still working) and are going to switch all the sites on the server onto a different one with Plesk 11 installed.
 
Back
Top