Description of problem: ======================= If volume options are modified when one of the peer node glusted is down, brick processes part of glusterd offline node are not getting the TCP ports once glusterd comes up. Version-Release number of selected component (if applicable): ============================================================= glusterfs-3.7.9-4.4 ( private build - glusterfs-3.7.9-4.4.gitb1ce6bc.el7rhgs.x86_64 ) How reproducible: ================= Always Steps to Reproduce: =================== 1.Create a sample volume using two node cluster bricks 2.Stop glusterd on one node (node1) 3.Modify any one of the volume options on node-2. 4.Start glusted on node1 and check the volume status. Actual results: ================ bricks processes not getting ports once glusterd comes up Expected results: ================= bricks processes should get the ports once glusterd comes up Additional info: will prove the logs.
We should not be filing bugs with private builds. If you hit the same issue on an official build please mention that otherwise you'd need to close it.
(In reply to Atin Mukherjee from comment #5) > We should not be filing bugs with private builds. If you hit the same issue > on an official build please mention that otherwise you'd need to close it. closing based on above comment as NOTABUG.
Clearing the flags so as this bug is not listed for bug triage
I am reopening this bug because issue found in the provide build is present in the official build "glusterfs-3.7.9-6" as well and setting the required flags back and this is regression b/w the builds "3.7.9-5" and "3.7.9-6" With official build i am seeing little diff in the issue. that is, only bricks part of node where glusterd was down and came up after modifying the volume options on other peer node are not getting the TCP ports. i will provide the required logs ( console and brick log failures ), these is no errors in glusterd log.
(In reply to Byreddy from comment #8) > I am reopening this bug because issue found in the private build is present > in the official build "glusterfs-3.7.9-6" as well and setting the required > flags back and this is regression b/w the builds "3.7.9-5" and "3.7.9-6" > > With official build i am seeing little diff in the issue. that is, only > bricks part of node where glusterd was down and came up after modifying the > volume options on other peer node are not getting the TCP ports. > > > i will provide the required logs ( console and brick log failures ), these > is no errors in glusterd log.
RCA: Since the downstream backport BZ 1335357 was not a straight forward one (refer to https://code.engineering.redhat.com/gerrit/74663 for more details on why), the patch changed the signature of glusterd_volinfo_copy_brick_portinfo () which lead to zeroing out brickinfo->port while importing a volume.
https://code.engineering.redhat.com/gerrit/74996 is posted for review.
https://code.engineering.redhat.com/gerrit/74996 is merged now.
Verified this bug using the build "glusterfs-3.7.9-7". Fix is working good, Moving to verified state.
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/RHBA-2016:1240