After an OCP upgrade, starting pod with previously defined block volume gives error "Heuristic determination of mount point failed - input/output error"
Issue
After an async OCP upgrade, provisioning new block storage works fine and also mounting these newly defined block volumes have no problems. But starting pod with previously defined block volume gives this error:-
MountVolume.WaitForAttach failed for volume "pvc-19bbe6df-2b73-11e9-b29d-005056a01756" : Heuristic determination of mount point failed:stat /var/lib/origin/openshift.local.volumes/plugins/kubernetes.io/iscsi/iface-default/xx.xxx.xxx.xxx:3260-iqn.2016-12.org.gluster-block:e85c4109-b7fe-484b-a678-c1e5756b03a4-lun-0: input/output error
Unable to mount volumes for pod "nginx-14-v72gf_jenkins(cc1230cb-411b-11e9-8e74-005056a039a3)": timeout expired waiting for volumes to attach or mount for pod "jenkins"/"nginx-14-v72gf". list of unmounted volumes=[volume-p5gbn]. list of unattached volumes=[nginx-1 nginx-build-trigger-x7v8d volume-p5gbn default-token-9wd6s]
An "ls" of /var/lib/origin/openshift.local.volumes/plugins/kubernetes.io/iscsi/iface-default/ shows I/O errors:-
[root@node01 ~]# cd /var/lib/origin/openshift.local.volumes/plugins/kubernetes.io/iscsi/iface-default/
[root@node01 iface-default]# ls -l
ls: cannot access xx.xxx.xxx.xxx:3260-iqn.2016-12.org.gluster-block:8f649bcc-71d6-419d-893c-f74b404505d2-lun-0: Input/output error
ls: cannot access xx.xxx.xxx.xxx:3260-iqn.2016-12.org.gluster-block:e85c4109-b7fe-484b-a678-c1e5756b03a4-lun-0: Input/output error
total 4
??????????? ? ? ? ? ? xx.xxx.xxx.xxx:3260-iqn.2016-12.org.gluster-block:e85c4109-b7fe-484b-a678-c1e5756b03a4-lun-0
??????????? ? ? ? ? ? xx.xxx.xxx.xxx:3260-iqn.2016-12.org.gluster-block:8f649bcc-71d6-419d-893c-f74b404505d2-lun-0
drwxrwsr-x. 19 root 1003130000 4096 Mar 13 21:17 xx.xxx.xxx.xxx:3260-iqn.2016-12.org.gluster-block:ea4ff328-1860-4041-8164-7d60b2244512-lun-0
Environment
OCP was upgraded from 3.10.89 to 3.10.111
No update of OCS was done
The pod that used the volume in question was force stopped
(oc delete pod --force --grace-period=0) as it did not stop gracefully
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.