Upgrade fails on Plesk Onyx Version 17.8.11 Update #7

[cyc]

Hi there,

we have a problem with the last Plesk upgrade. First time the upgrade failed because of a missing WordpressInstance table. No idea why it was missing, but we did find this link and created the table, that now looks like:

MariaDB [psa]> SHOW COLUMNS FROM WordpressInstances;
+----------------+------------------+------+-----+---------+----------------+
| Field | Type | Null | Key | Default | Extra |
+----------------+------------------+------+-----+---------+----------------+
| id | int(10) unsigned | NO | PRI | NULL | auto_increment |
| subscriptionId | int(10) unsigned | NO | MUL | NULL | |
| path | varchar(255) | NO | | NULL | |
| isIgnored | int(11) | NO | | 0 | |
| apsInstanceId | int(11) unsigned | YES | | NULL | |
+----------------+------------------+------+-----+---------+----------------+
5 rows in set (0.00 sec)

After creating that table we tried to update Plesk again (Tools & Settings -> Updates and Upgrades -> Update Components -> Plesk). Again we get an error for a missing table, this time for ServiceInstances:

ERROR 1146 (42S02) at line 1: Table 'psa.ServiceInstances' doesn't exist
...
ERROR while trying to execute SQL query, the query was: ALTER TABLE `ServiceInstances` ADD UNIQUE `reference` (serviceInstallationId,`servicePackageId`,`reference`)
(of course, the table really does seem to be missing in psa database)

For tha last table we cannot find any "create table" ... can someone please help?

 

[IgorG]

What about just

# plesk repair db

?

 

[cyc]

Sadly no - that's the first thing I tried and it didn't help. That suggestion is on the link I posted as well and then it didn't help either.

I wanted to paste that output showing everything is fine, but now I get a warning about an extra table:

***
Checking the Plesk database using the native database server tools .. [OK]

Checking the structure of the Plesk database

Extra table: WordpressInstances ................................... [WARNING]
Delete table 'WordpressInstances'? [Y/n]
***

Of course, I didn't create that table because I was bored, Plesk installer failed because of the missing table. After that table was created, it fails because of another missing table. But yes, "plesk repair db" claims there's an extra table there. This was from the first try:

ERROR 1146 (42S02) at line 1: Table 'psa.WordpressInstances' doesn't exist

Now what ...? It's "extra" but it'll be missed after deleting ...

 

[cyc]

I think I've found & fixed the problem. "apt-get upgrade" showed a possible update of various plesk-* packages:

***
The following packages will be upgraded:
plesk-repair-kit plesk-service-node-utilities plesk-skins psa-locale-base-en-us
4 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
***

After the update, component update were trying to update went through successfully. "plesk repair db" found two other problems and now they're all fixed now:

Checking the Plesk database using the native database server tools .. [OK]

Checking the structure of the Plesk database

Extra table: WordpressInstances ................................... [WARNING]
Delete table 'WordpressInstances'? [Y/n]
Creating a database dump ........................................ [OK]
A dump of the database was created before an attempt to repair the database: /var/lib/psa/dumps/mysql.preresolve.20180601-135001.dump.gz
Deleting table 'WordpressInstances' ............................. [FIXED]

The table EmailActivations is absent .............................. [ERROR]
Create table 'EmailActivations'? [Y/n]
Creating table 'EmailActivations' ............................... [FIXED]

Column isActivationRequired is absent in the table smb_roles ...... [ERROR]
Repair table 'smb_roles'? [Y/n]
Repairing table 'smb_roles' ..................................... [FIXED]

Checking the consistency of the Plesk database ...................... [OK]

Error messages: 2; Warnings: 1; Errors resolved: 3

Our problem seems to be solved. I don't understand what caused it, maybe partially downloaded/installed plesk packages ...

 

[cyc]

Hi there,

the same error repeated today. It stated with a nice mail from Plesk telling us "Plesk update failed" during the night. Sure enough, in /var/log/plesk/install/plesk_17.8.11_installation.log there is this error again:

ERROR 1146 (42S02) at line 1: Table 'psa.WordpressInstances' doesn't exist

... and the log ends with:

----------------------------
END OF INNODB MONITOR OUTPUT
============================


ERROR while trying to execute SQL query, the query was: ALTER TABLE `WordpressInstances` ADD KEY `subscriptionId_idx` (subscriptionId)
Check the error reason(see log file: /var/log/plesk/install/plesk_17.8.11_installation.log), fix and try again


===> Restoring database from backup /var/lib/psa/dumps/mysql.preupgrade.17.8.11-17.8.11.20180925-152455.dump.gz


 Trying to start psa...
Stopping sw_engine service... already stopped

**** Product prep-install for BASE completed with warnings.

*****  problem report *****
ERROR while trying to execute SQL query, the query was: ALTER TABLE `WordpressInstances` ADD KEY `subscriptionId_idx` (subscriptionId)
Check the error reason(see log file: /var/log/plesk/install/plesk_17.8.11_installation.log), fix and try again
STOP Bootstrapper 17.8.11 prep-install for BASE AT Tue Sep 25 15:25:07 CEST 2018

By the way, isn't the "Check the error reason(see log file: /var/log/plesk/install/plesk_17.8.11_installation.log), fix and try again" line a bit silly considering it's taken from /var/log/plesk/install/plesk_17.8.11_installation.log?

Because of previous experiences with the same error, I executed "apt-get update" & "apt-get upgrade" and the following two packages were updated:

Inst psa-locale-base-en-us [17.8.11-debian9.0.build1708180823.15] (17.8.11-debian9.0.build1708180912.14 Plesk 17.8.11 packages:stretch [all])
Inst psa-phpmyadmin [4.7.8-debian9.0.build1708180222.15] (4.8.3-debian9.0.build1708180919.15 Plesk 17.8.11 packages:stretch [all])

After the update, manual execution of Plesk's update went through fine. But I'm starting to see a pattern here ... I have no clue what causes these "hiccups", but I'm sure they don't happen on every update of Plesk. Do we need to change anything or lies the problem elsewhere?

I forgot to mention - output of "repair plesk db" was fine before - when update failed - and now after the Plesk update:

Checking the Plesk database using the native database server tools .. [OK]

Checking the structure of the Plesk database ........................ [OK]

Checking the consistency of the Plesk database ...................... [OK]

Error messages: 0; Warnings: 0; Errors resolved: 0

 

[IgorG]

ERROR while trying to execute SQL query, the query was: ALTER TABLE `WordpressInstances` ADD KEY `subscriptionId_idx` (subscriptionId) Check the error reason(see log file: /var/log/plesk/install/plesk_17.8.11_installation.log), fix and try again

Please check this KB article Unable to update Plesk: ERROR while trying to execute SQL query

 

[cyc]

Hello again,

upgrader_config_current_timestamp contained value: 20140511180000
I changed it to 20180301000200, because there's Onyx 17.8.11 on the host. The last past of your KB worries me "a bit". Especially when I saw the "New installation of product(s) should be performed on clean servers only" line. That host isn't empty at all. Will I run into problems if I continue?

The following product versions are available:

1. [*] Plesk
2. (*) Plesk Onyx 17.8.11 (Stable) (currently installed)