Bug 50277 - Don't handle objects in ou=temporary_move_container_*
Don't handle objects in ou=temporary_move_container_*
Status: NEW
Product: UCS
Classification: Unclassified
Component: S4 Connector
UCS 4.4
Other Linux
: P5 normal (vote)
: ---
Assigned To: Samba maintainers
Samba maintainers
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2019-09-26 15:50 CEST by Florian Best
Modified: 2021-03-17 08:52 CET (History)
1 user (show)

See Also:
What kind of report is it?: Development Internal
What type of bug is this?: ---
Who will be affected by this bug?: ---
How will those affected feel about the bug?: ---
User Pain:
Enterprise Customer affected?: Yes
School Customer affected?:
ISV affected?:
Waiting Support:
Flags outvoted (downgraded) after PO Review:
Ticket number:
Bug group (optional):
Max CVSS v3 score:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Florian Best univentionstaff 2019-09-26 15:50:50 CEST
As can be seen in Bug #49876 comment 0:
> noObject: uid=Xprdhkm2sxk,cn=q5x151o5zv,ou=temporary_move_container_1562854465.1,dc=AutoTest091,dc=local

The S4-Connector modifies objects which are underneath of "ou=temporary_move_container_*" containers.
This container is a temporary OU which exists for a short period during renaming containers.

This is very error prone, as some changes probably don't get applied due to race conditions: 7: causes data lost.

I think the S4-Connector should not touch currently being moved objects and wait until these objects are moved.
Maybe the objects where only touched because they were detected as moved(?).

My suggestion would be to not write 2 pickle files in the s4connector listener with the interim state (if new_dn or old_dn starts with ou=temporary_move_container_*).
Instead this state should be detected and only one pickle file should be created after the new location is known.

The S4-Connector should also know that there is currently a move (container rename) operation in progress and all changes which are done on Samba side should be delayed after the move operation succeeded.
Comment 1 Marc Schwarz univentionstaff 2021-03-08 08:57:19 CET
I noticed a very similar behaviour in an environment, where the AD-Conenctor is used. There are always the same 3 or 4 users affected.
Comment 2 Florian Best univentionstaff 2021-03-16 21:50:15 CET
(In reply to Marc Schwarz from comment #1)
> I noticed a very similar behaviour in an environment, where the AD-Conenctor
> is used. There are always the same 3 or 4 users affected.

It would probably be good if you describe this a little bit further, attach a ticket number, set the customer ID and Enterprise-Customer affected.
Comment 3 Marc Schwarz univentionstaff 2021-03-17 08:46:23 CET
sorry, i wasn't able to get the logs in time; just made my note here first, so that I don't forget to attach them.

I added customer-ID, there is no OTRS ticket yet. Logs will follow.