Storage repair iterates over all LDAP users during upgrade

[sbraz]

As discussed with @DeepDiver1975, while upgrading from 9.1.0 to 9.1.2 for an instance which uses LDAP, the code iterates over all users matched by the LDAP filter, not only those who connected once.

[DeepDiver1975]

https://github.com/owncloud/core/blob/master/lib/private/Repair/RepairLegacyStorages.php#L212

we should iternate only over the known users

[PVince81]

I just realized a flaw of this: if ($this->config->getAppValue('core', 'repairlegacystoragesdone') === 'yes').

This repair was supposed to run only for installs that were created before OC 8.1 or so, then never run again thanks to this flag.

However this flag isn't set for new OC >= 8.1 new installs, so it would run at least for the very first upgrade. This is wrong.

One idea would be to kill that repair step completely in OC >= 8.1.

[PVince81]

Quick workaround:
insert into oc_appconfig values ('core', 'repairlegacystoragesdone', 'yes');

[PVince81]

From what I see that class was introduced in OC 8.0, so it's the upgrade from OC 7 to OC 8 that needed this repair. If the repair has been done already then there is no need to do it in later versions.

[PVince81]

One commonly observed problem is that on some setups people ignored the warnings given by this repair step in case of unrepairable situations. Then the duplicate storages still exist up to current versions. (the duplicates could only have been created in OC < 8.0).

But with that reasoning, if the duplicates could not be fixed automatically, the repair step itself can still be removed from the codebase as it wouldn't be useful anyway in such situations.

However, the legacy storage fallback would still make sure that OC still works as before. There are discussions about removing the fallback as well, which is a much more risky change: #26325

[PVince81]

PR to remove the repair step completely #26779

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.