Hide Forgot
Description of problem: On 3 capsules qdrouterd was killed by oom-killer after 2k (2k on each, 6k in total) clients got errata applied. Version-Release number of selected component (if applicable): Satellite 6.2.1 How reproducible: always on our setup Steps to Reproduce: 1. Register 6k clients via 3 capsules 2. Apply errata on them Actual results: After some time, qdrouterd on capsules gets killed Expected results: qdrouterd keeps only memory it requires
There were various leaks and multiple BZs that we believe are hitting the same issues. For router, we think this (1369835) and 1368718 have been covered by the fix for 1358948. This BZ was originally from 6.2.1, while the latest qpid-dispatch (0.4-21) was shipped in December errata https://access.redhat.com/errata/RHBA-2016:2855. The last scale test showed some bad results, but the bad capsules had the wrong packages installed. We'd like to see testing with all the latest packages and proceed from there.
(In reply to Pavel Moravec from comment #28) > .. .. and that resulted into bz1523793 . When now reviewing setup for this BZ, I tend to state both the BZs describe the same underlying bug. Since here: - different qdrouterd versions were used, causing the routers were not able to communicate / route messages or link routing requests - that is another way of triggering 'qd:no-route-to-dest' or another link error that the router on Capsule sends to the client While bz1523793 is triggered by inability of proper link routing towards qpidd while (goferd) clients are repeatedly requesting those links. Since not having any better reproducer or explanation here, I suggest closing either this or the other BZ as a duplicate of the other one.
Verified in Satellite 6.3 Snap 35. Based on comment https://bugzilla.redhat.com/show_bug.cgi?id=1523793#c5, this has been fixed in qpid-dispatch-router-0.4-28+ Currently packaged version in Sat 6.3 Snap 35 is qpid-dispatch-router-0.8.0-16. Additionally, there have been no observed breakages during testing.
*** Bug 1523793 has been marked as a duplicate of this bug. ***
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. > > For information on the advisory, and where to find the updated files, follow the link below. > > If the solution does not work for you, open a new bug report. > > https://access.redhat.com/errata/RHSA-2018:0336