Description of problem: ======================= Getting the below warning messages in glusterd and volume mount logs continuously on SSL SETUP when one of the cluster node glusterd is down/node is down. Message frequency is very high and if cluster node is down for any known reason for days, it will consume lot of log storage. [2016-10-17 08:15:55.008898] W [socket.c:590:__socket_rwv] 0-management: readv on 10.70.43.190:24007 failed (No data available) [2016-10-17 08:15:55.009082] W [socket.c:590:__socket_rwv] 0-management: readv on 10.70.43.190:24007 failed (No data available) [2016-10-17 08:15:55.009289] W [socket.c:590:__socket_rwv] 0-management: readv on 10.70.43.190:24007 failed (No data available) [2016-10-17 08:15:55.009546] W [socket.c:590:__socket_rwv] 0-management: readv on 10.70.43.190:24007 failed (No data available) [2016-10-17 08:15:55.009736] W [socket.c:590:__socket_rwv] 0-management: readv on 10.70.43.190:24007 failed (No data available) Version-Release number of selected component (if applicable): ============================================================= glusterfs-3.8.4-2 How reproducible: ================= Always Steps to Reproduce: =================== 1. Have two node cluster 2. Create a SSL configuration. 3. Create and start any volume type 4. fuse mount the volume. 5. Stop glusterd on vol file server node/ shutdown the node 6. Check the glusterd logs in the peer nodes and the volume mount log for the above mentioned warning messages Actual results: =============== Continuous warning messages getting when one of the cluster node is down on SSL setup. Expected results: ================= There should be control on throwing the warning messages or it should not throw the warning messages. Additional info: ================ I used backup vol file server while mounting the volume then also issue remains same in the volume mount log.
This issue is not there in the last GA build.
Hi, At the time of start glusterd process it returns a port to the other peer node on which it will communicate with glusterd(24007).After stop glusterd on a node means one end point of port is closed and other point is still used by socket(glusterd 24007),glusterd(socket_poller) on other node call's continuously socket_event_poll_in that calls __socket_rwv to read data from socket unless buffer is not clear. Because other end point of socket is disconnected so it prints the message "0-management: readv on 10.65.7.252:24007 failed (No data available)". After changed the condition to log the message issue will resolve. Regards Mohit Agrawal
upstream mainline : http://review.gluster.org/15677 upstream 3.9 : http://review.gluster.org/15711
I am seeing little diff in the issue with the build 3.8.4-3 with same scenario, updating the result here. Getting the same readv messages in the mount log for the above scenarios mentioned BUT in peer nodes glusterd log, getting the below error messages. [2016-10-28 09:03:31.016907] E [socket.c:3097:socket_connect] 0-management: connection attempt on 10.70.41.198:24007 failed, (Connection refused) [2016-10-28 09:03:31.017272] E [socket.c:353:ssl_setup_connection] 0-management: SSL connect error (client: 10.70.41.198:24007) [2016-10-28 09:03:31.017368] E [socket.c:2436:socket_poller] 0-management: client setup failed [2016-10-28 09:03:34.026491] E [socket.c:3097:socket_connect] 0-management: connection attempt on 10.70.41.198:24007 failed, (Connection refused) [2016-10-28 09:03:34.026861] E [socket.c:353:ssl_setup_connection] 0-management: SSL connect error (client: 10.70.41.198:24007) [2016-10-28 09:03:34.027166] E [socket.c:2436:socket_poller] 0-management: client setup failed [2016-10-28 09:03:37.035856] E [socket.c:3097:socket_connect] 0-management: connection attempt on 10.70.41.198:24007 failed, (Connection refused) [2016-10-28 09:03:37.036416] E [socket.c:353:ssl_setup_connection] 0-management: SSL connect error (client: 10.70.41.198:24007) [2016-10-28 09:03:37.036545] E [socket.c:2436:socket_poller] 0-management: client setup failed [2016-10-28 09:03:40.045129] E [socket.c:3097:socket_connect] 0-management: connection attempt on 10.70.41.198:24007 failed, (Connection refused) [2016-10-28 09:03:40.045473] E [socket.c:353:ssl_setup_connection] 0-management: SSL connect error (client: 10.70.41.198:24007) in the previous build i was getting readv messages in peer nodes glusterd log and in volume mount log.
downstream patch : https://code.engineering.redhat.com/gerrit/#/c/89220/
Verified this bug using glusterfs-3.8.4-5 build. Fix is working good. Reported issue is not seen with this build 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://rhn.redhat.com/errata/RHSA-2017-0486.html