Resolved Unable to run plesk commands

[Brian Taylor]

I have been away from this server for over a month and come back to it to find I cannot access Plesk in any form.

• The control panel just gives me 'Unable to connect'
• Typing into the command line, gives no response other than giving me the command prompt

The website that is run on the server is ok and working fine.

On looking through the logs and checking backups.

• The last backup was on the 2nd of November.
• Various error messages indicating file too short or no such file. Examples
  • PAM unable to dlopen(/usr/lib64/security/pam_plesk.so): /usr/lib64/security/pam_plesk.so: file too short
  • "/usr/local/psa/admin/htdocs/currentsetting.htm" failed (2: No such file or directory),
• When I look in various plesk directories, almost every file is listed as 0 filesize, see examples attached tmp80.jpeg - tmp82.jpeg
• Looking in the installation logs, plesk_18.3.31_installation.log was created on 2nd Nov but is 0 filesize

Everything seems to be pointing to a failed upgrade. I am not finding a way to re trigger the upgrade or restore the server back. Probably a failing in my understanding of what the plesk backups store, in asking for a full system backup, that was what it does. Maybe need to look into a different way to secure the system. Having said that, I cannot run the Plesk Restore backup command to restore it anyway. No response when I type in any plesk commands see attached tmp83.jpeg

Server setup is
Centos 7.8.2003
Apache 2.4.6
Php 7.2
Plesk Obsidian 18.0.30

Looking for a way to either

• restore from the plesk backups,
• rerun the failed upgrade to find the problem
• reinstall plesk
• download Plesk and reapply missing files.

Any and all suggestions welcome, if further information needed just ask and I will attempt to get answers.

Thanks

 

[Bitpalast]

Do you see entries in /var/log/plesk/install/autoinstaller3.log that point to a failed upgrade?

 

[Brian Taylor]

Yes, looks like it. last entries are
Running Transaction Check
Updating: 1lesk-php73-7.3.24-1centos.7.201030.1238.x86_64 [1/250]
Updating: plesk-php74-7.4.12-1centos.7.201030.1238.x86_64 [2/250]
Updating: 1lesk-php73-pdo-7.3.24-1centos.7.201030.1238.x86_64 [3/250]
Updating: plesk-php74-pdo-7.4.12-1centos.7.201030.1238.x86_64 [4/250]
Updating: 1lesk-php73-xml-7.3.24-1centos.7.201030.1238.x86_64 [5/250]
Updating: 1lesk-php73-cli-7.3.24-1centos.7.201030.1238.x86_64 [6/250]
Updating: plesk-dovecot-core-2.3.10.1-2.centos.7+p18.0.31.0+t201022.1604.x86_64 [7/250]
Updating: 1lesk-php73-mysql-7.3.24-1centos.7.201030.1238.x86_64 [8/250]
Updating: 1lesk-php73-imap-7.3.24-1centos.7.201030.1238.x86_64 [9/250]
Updating: 1lesk-php73-mbstring-7.3.24-1centos.7.201030.1238.x86_64 [10/250]
Updating: sw-librrd-1.6.0.1-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [11/250]
Updating: psa-mod_fcgid-2.3.9.5-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [12/250]
Updating: plesk-platform-runtime-1.0.2-1centos.7.200921.1614.x86_64 [13/250]
Updating: psa-mod_proxy-2.4.6-2.centos.7+p18.0.31.0+t200930.1833.x86_64 [14/250]
Updating: plesk-ui-library-18.0-2.centos.7+p18.0.31.0+t201022.2113.noarch [15/250]
Updating: plesk-control-panel-18.0-2.centos.7+p18.0.31.0+t201022.2113.noarch [16/250]
Updating: plesk-rdbmspp-2.0.2-1centos.7.200921.1614.x86_64 [17/250]
Updating: libaps-1.0.10-1centos.7.200921.1614.x86_64 [18/250]
Updating: sw-collectd-5.11.0-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [19/250]
Warning: sw-collectd.service changed on disk. Run 'systemctl daemon-reload' to reload units.
Updating: plesk-php74-cli-7.4.12-1centos.7.201030.1238.x86_64 [20/250]
Updating: plesk-php74-xml-7.4.12-1centos.7.201030.1238.x86_64 [21/250]
Updating: 1lesk-php73-gd-7.3.24-1centos.7.201030.1238.x86_64 [22/250]
Updating: plesk-release-18.0.31.0-2.centos.7+p18.0.31.0+t201025.0009.x86_64 [23/250]
Updating: plesk-libmaodbc-3.0.8-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [24/250]
Updating: log4cplus-1.2.0.1-1centos.7.200921.1614.x86_64 [25/250]
Updating: plesk-lmlib-0.2.4-1centos.7.200921.1614.x86_64 [26/250]
Updating: sw-engine-2.31.0-1centos.7.200921.1614.x86_64 [27/250]
Updating: 1:fail2ban-0.10.3.1-2.centos.7+p18.0.31.0+t201020.1812.noarch [28/250]
Warning: fail2ban.service changed on disk. Run 'systemctl daemon-reload' to reload units.
Updating: sw-cp-server-2.8.0-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [29/250]
Updating: plesk-engine-18.0-2.centos.7+p18.0.31.0+t200930.1833.noarch [30/250]
Updating: psa-18.0.31-2.centos.7+p18.0.31.0+t201022.1604.x86_64 [31/250]
Updating: plesk-dovecot-2.3.10.1-2.centos.7+p18.0.31.0+t201022.1604.x86_64 [32/250]
Updating: plesk-dovecot-pigeonhole-0.5.10-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [33/250]
Updating: psa-proftpd-1.3.6c-2.centos.7+p18.0.31.0+t201022.1604.x86_64 [34/250]
Updating: psa-locale-base-en-US-18.0-2.centos.7+p18.0.31.0+t201025.0009.noarch [35/250]
Updating: plesk-completion-18.0-2.centos.7+p18.0.31.0+t201005.1334.x86_64 [36/250]
Updating: plesk-core-utilities-18.0-2.centos.7+p18.0.31.0+t201023.1535.x86_64 [37/250]
Updating: psa-libxml-proxy-2.9.7-2.centos.7+p18.0.31.0+t201020.1812.x86_64 [38/250]
Updating: plesk-backup-utilities-18.0-2.centos.7+p18.0.31.0+t201023.1535.x86_64 [39/250]
Updating: psa-libpam-plesk-18.0-2.centos.7+p18.0.31.0+t201022.1604.x86_64 [40/250]
Updating: psa-phpmyadmin-5.0.3-2.centos.7+p18.0.31.0+t201014.1920.noarch [41/250]

 

[Bitpalast]

The log simply stops at that point without any further entries?
What does /var/log/messages say at the same time?

 

[Brian Taylor]

Yes stops there, nothing further after that.
/var/log/messages doesn't go back far enough, the oldest one I've got is 20201122. Whatever error there was from the 2nd, I have no idea

 

[Bitpalast]

I'd guess it is a hard disk issue. Either the disk was full or another issue with the file system occured. It's probably not been a software problem, because in that case there would have been further log entries. If it is a virtual server, the error could have been caused by an issue not in your container, but on the physical system, too.

Check if you have enough free disk space:
# df

Check your file system. First determine what you got
# parted /dev/sda 'print'
Then use fsck to check these partitions, example:
fsck -M -t <file system type> -f -n <device name>
examples:
fsck -M -t ext3 -f -n /dev/sda2
fsck -M -t ext4 -f -n /dev/sda4
File system type must be correct and match the devices file system, else a lot of damage can be caused.

 

[Brian Taylor]

See attached. df shows there should be space available.
parted shows the partition uses lvm, is it ok to use fsck to check further?

 

[Bitpalast]

I cannot assist you with this further, because I don't know enough about lvm on top of an old msdos partition (as it looks to me). Not sure what damage could be caused or what that situation is. You'll need to ask someone who is an expert for your system or your OS.

 

[Brian Taylor]

Thanks for having a look. At least I know for certain the upgrade was interrupted, just need to find the cause.

 

[Brian Taylor]

Checking through for space issues, everything looks ok for that

 

[Brian Taylor]

Talking to my server provider, they found a warning message in the upgrade log

[2020-11-02 09:01:02.225576] Warning: Failed to attach to cgroup /sys/fs/cgroup/systemd/user/root/tasks, installer might be unexpectedly killed later by the system

and pointed out a help article related to this warning on the Plesk support page

Plesk upgrade/update fails: Error: Failed to attach to cgroup

Applicable to: Plesk 12.0 for Linux Plesk Onyx for Linux Symptoms Plesk update/upgrade fails with the following error in /var/log/plesk/install/autoinstaller3.log file: CONFIG_TEXT: [2017-0...

support.plesk.com

They tried to rerun the upgrade, which got further but still failed further along the process. This is being passed to Plesk Support to see if they can help

 

[Brian Taylor]

This has now been resolved by Plesk Support, Copy of the transcript saying what they needed to do.

This is the steps that have taken to resolve the issue

I found out that Plesk database could not be extracted, I got some warning in logs:

2020-12-17 23:53:40 140071953798912 [Warning] InnoDB: Tablespace 'psa/clients' exists in the cache with id 401484 != 401694

So I deleted psa database:

mysql> drop database psa;

Then I restarted Mariadb and restored database from backup.

Then I found out that sw-cp-service does not start:

# service sw-cp-server start
Starting sw-cp-server (via systemctl): Job for sw-cp-server.service failed because the control process exited with error code. See "systemctl status sw-cp-server.service" and "journalctl -xe" for details.
[FAILED]

because /etc/sw-cp-server/config is empty:

[root@kingdom-sales ~]# ls -la /etc/sw-cp-server/config
-rw-r--r--. 1 root root 0 Oct 20 12:18 /etc/sw-cp-server/config

So I reinstalled sw-cp-server package:

[root@kingdom-sales ~]# rpm -Uvh http://autoinstall.plesk.com/PSA_18.0.32/dist-rpm-CentOS-7-x86_64/engine/sw-cp-server-2.8.0-2.centos.7+p18.0.32.1+t201211.1937.x86_64.rpm --force

Now all Plesk services are running. Then I found out that "plesk installer" command does not work, because there was empty file /usr/local/psa/version.upg.​ So I removed these file and launched plesk installer without any issue. Plesk works now as well as website

Have to thank the support of Easyspace and Plesk in their prompt response and in sorting out my problem.

 

[Bitpalast]

2020-12-17 23:53:40 140071953798912 [Warning] InnoDB: Tablespace 'psa/clients' exists in the cache with id 401484 != 401694

...

because /etc/sw-cp-server/config is empty:

...

there was empty file /usr/local/psa/version.upg.​

All three pointing to a hard disk failure again, specifically the last entry, because the file name is faulty. Maybe you have a chance to check your disk to avoid similar issues in the future.