Bug 24761 - Caching issues in scripts running in daemon mode [OMNIBUS]
Summary: Caching issues in scripts running in daemon mode [OMNIBUS]
Status: NEW
Alias: None
Product: Koha
Classification: Unclassified
Component: Architecture, internals, and plumbing (show other bugs)
Version: master
Hardware: All All
: P5 - low normal (vote)
Assignee: Bugs List
QA Contact: Testopia
URL:
Keywords:
Depends on: 16758 23640 25292
Blocks:
  Show dependency treegraph
 
Reported: 2020-02-28 14:25 UTC by Tomás Cohen Arazi
Modified: 2020-04-28 00:49 UTC (History)
5 users (show)

See Also:
Change sponsored?: ---
Patch complexity: ---
Documentation contact:
Documentation submission:
Text to go in the release notes:
Version(s) released in:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Tomás Cohen Arazi 2020-02-28 14:25:35 UTC
Bug 16758 highlighted an issue on long running processes (daemons) with the L1 cache mechanism not being able to invalidate cache.

The proposed and agreed solution there was to disable L1 cache for daemon scripts. But the bug only took care of rebuild_zebra.pl, leaving things like SIPServer for later. We should resume that work, and discussion.

It clearly needs to be handled in a generic way on Koha::Script
Comment 1 Jonathan Druart 2020-02-28 15:02:43 UTC
Invalidation of L1 cache for SIP has been done on bug 23640.
Comment 2 Jonathan Druart 2020-04-27 14:06:50 UTC
Opened Bug 25292 - L1 cache too long in Z3950 server (z3950-responder)

Do we have something else?
Comment 3 David Cook 2020-04-28 00:48:05 UTC
> Do we have something else?

I can't think of anything else?

In terms of Koha-related daemons, at a glance all I see are the following:
- koha-plack (starman)
- koha-indexer (rebuild_zebra.pl)
- koha-sip (SIPServer.pm)
- koha-zebra (zebrasrv)
- koha-z3950-responder (Z3950Responder.pm)

koha-plack and koha-zebra are irrelevant. koha-indexer and koha-sip are done. So that should just leave koha-z3950-responder via https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=25292.
Comment 4 David Cook 2020-04-28 00:49:26 UTC
(In reply to Jonathan Druart from comment #2)
> Do we have something else?

If Bug 22417 goes ahead, the koha-worker daemons would probably need it, if they were closely coupled to Koha.