|
Bugzilla – Full Text Bug Listing |
| Summary: | mandb.service: "find: ‘standard output’: Broken pipe", "find: write error", "status=0/SUCCESS" | ||
|---|---|---|---|
| Product: | [openSUSE] openSUSE Distribution | Reporter: | Martin Jacobs <martin.jacobs> |
| Component: | Basesystem | Assignee: | Dr. Werner Fink <werner> |
| Status: | RESOLVED FIXED | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Normal | ||
| Priority: | P5 - None | CC: | alynx.zhou, martin.jacobs, qa-bugs, Ulrich.Windl, werner |
| Version: | Leap 15.4 | ||
| Target Milestone: | --- | ||
| Hardware: | Other | ||
| OS: | openSUSE Leap 15.4 | ||
| Whiteboard: | |||
| Found By: | --- | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Bug Depends on: | 1155879 | ||
| Bug Blocks: | |||
| Attachments: |
Output of sudo bash -x /usr/lib/man-db/do_mandb |& tee do_mandb.log
/usr/lib/man-db/do_mandb |
||
|
Description
Martin Jacobs
2023-07-23 16:48:40 UTC
Run as root bash -x /usr/lib/man-db/do_mandb |& tee do_mandb.log and attach do_mandb.log Btw: Are you using bash or an other /bin/sh ... report realpath /bin/sh (In reply to Dr. Werner Fink from comment #2) > Btw: Are you using bash or an other /bin/sh ... report > > realpath /bin/sh Simple answer first: result of realpath /usr/bin/bash Created attachment 868403 [details] Output of sudo bash -x /usr/lib/man-db/do_mandb |& tee do_mandb.log Answer to comment #1 (In reply to Martin Jacobs from comment #4) > Created attachment 868403 [details] > Output of sudo bash -x /usr/lib/man-db/do_mandb |& tee do_mandb.log > > Answer to comment #1 Hmm ... no errors ... also here mandb.service works flawsless ... please edit /usr/lib/man-db/do_mandb and change the veriy first line by adding a space and -x to the /bin/sh ... that is #!/bin/sh becomes #!/bin/sh -x and then to systemctl start mandb.service to see what had happen please call journalctl -b 0 -o with-unit --unit mandb.service > mandb.log and attach the file mandb.log Maybe I did forget to mention: The error seems to occur only as part of system start, while many tasks run in parallel. I did recognize this only because command apropos did not give any results.
Here are top 5 lines of output of systemd-analyze blame
1min 8.819s mandb.service
1min 8.238s mlocate.service
26.635s clamd.service
23.526s backup-rpmdb.service
21.007s logrotate.service
The astonishing result is, that mandb.service is listed in top, even before mlocate.service and clamd.service!
mandb.service is NOT part of critical chain!
I guess the performance issue of mandb.service triggers timeouts on read statement (-t 1) and thus hinders mandb.service to successfully complete its job.
And running do_mandb from the command line after boot does not give more insight, as it runs fine in this state if the system.
Created attachment 868432 [details] /usr/lib/man-db/do_mandb (In reply to Martin Jacobs from comment #6) > Maybe I did forget to mention: The error seems to occur only as part of > system start, while many tasks run in parallel. I did recognize this only > because command apropos did not give any results. > > Here are top 5 lines of output of systemd-analyze blame > > 1min 8.819s mandb.service > 1min 8.238s mlocate.service > 26.635s clamd.service > 23.526s backup-rpmdb.service > 21.007s logrotate.service > > The astonishing result is, that mandb.service is listed in top, even before > mlocate.service and clamd.service! > > mandb.service is NOT part of critical chain! > > > I guess the performance issue of mandb.service triggers timeouts on read > statement (-t 1) and thus hinders mandb.service to successfully complete its > job. > > And running do_mandb from the command line after boot does not give more > insight, as it runs fine in this state if the system. Please test out the attachment as replacement of /usr/lib/man-db/do_mandb but make sure to run chmod a+x /usr/lib/man-db/do_mandb And here's result of todays boot:
top 5 lines from systemd-analyze blame
1min 32.931s mlocate.service
1min 13.890s mandb.service
30.181s clamd.service
24.693s backup-rpmdb.service
13.217s dev-disk-by\x2duuid-a988af6f\x2de4bd\x2d4e00\x2dab22\x2d5092449311ee.device
And result of journalctl -b 0 -o with-unit --unit mandb.service
Thu 2023-07-27 20:17:33 CEST Schnecke2G.Jacobs.home init.scope[1]: Starting Do daily mandb update...
Thu 2023-07-27 20:18:48 CEST Schnecke2G.Jacobs.home init.scope[1]: mandb.service: Deactivated successfully.
Thu 2023-07-27 20:18:48 CEST Schnecke2G.Jacobs.home init.scope[1]: Finished Do daily mandb update.
So, your patch improves behavior of do_mandb definitely!
Same result today. :-) Now systemd timer based |