| Summary: | subscription-manager has to restart dbus after it change definition in /etc/dbus-1/system.d/com.redhat.SubscriptionManager.conf | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Miroslav Suchý <msuchy> |
| Component: | subscription-manager | Assignee: | Chris Duryee <cduryee> |
| Status: | CLOSED WORKSFORME | QA Contact: | IDM QE LIST <seceng-idm-qe-list> |
| Severity: | high | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 6.2 | CC: | bkearney, cduryee |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2012-03-27 17:12:05 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Bug Depends On: | |||
| Bug Blocks: | 738066 | ||
|
Description
Miroslav Suchý
2011-12-20 12:45:39 UTC
I just checked rhel5 and it does not seem to be affected by this problem. Thanks for the detailed bug report. This appears to be an issue with dbus not reloading its config without restart. In %post and %preun, we do the following: dbus-send --system --type=method_call --dest=org.freedesktop.DBus / org.freedesktop.DBus.ReloadConfig This should reload dbus configs. However, I am able to repro this issue and when I run the above command, it does not reload the dbus config. It appears that only a service messagebus restart does the job. I will look more into this to see if there is a better way to restart dbus. Miroslav, It appears that after a few seconds, dbus will reload the config on its own. I wrote two small test scripts, test-61.py which uses the Compliance dbus call, and test-62.py which uses EntitlementStatus. Both scripts print which version of the call they're making, and then attempt to make the call. I updated susbcription-manager, and then the following four commands were run within seconds of each other: [root@six-one-up ~]# python test-62.py ENTITLEMENT <error> [root@six-one-up ~]# python test-61.py COMPLIANCE [root@six-one-up ~]# python test-61.py COMPLIANCE <error> [root@six-one-up ~]# python test-62.py ENTITLEMENT Do you see the same behavior? No. I just installed rhel6.0, run script from #0 which fail with TB from #0 (this is expected). then upgraded to subscription-manager-0.96.17-1.el6.x86_64. Run script again and I still got TB. I tried to run it in different intervals for 7 minutes and still no luck. I even manualy run dbus-send --system --type=method_call --dest=org.freedesktop.DBus / org.freedesktop.DBus.ReloadConfig still the same TB. Once I executed: service messagebus restart The script from #0 run without problem. Miroslav, I was installed rhel 6.0 on a vm and did the same upgrade procedure, but I still get the same behavior mentioned in comment #4. Can you do the same test as you did in comment #5, but instead of performing the dbus-send command manually, do the following instead: kill -s SIGHUP `cat /var/run/messagebus.pid` If that works, I will swap the dbus-send call for the SIGHUP call in the spec file, and we can perform testing to ensure it doesn't create any regressions. Miroslav, I was unable to repro this, and awood was unable as well. We spoke to jsefler, and given the severity of the bug and our inability to reproduce it on clean installs, we agreed that it was ok to close as WORKSFORME. If this bug becomes more of an issue, let us know and we will look into it some more. |