Bug 1669403 - With one target node down, on respinning the app pod, block device fails to mount
Summary: With one target node down, on respinning the app pod, block device fails to m...
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: kubernetes
Version: ocs-3.11
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
: OCS 3.11.z Batch Update 3
Assignee: Humble Chirammal
QA Contact: Rachael
URL:
Whiteboard:
Depends On: 1669560 1680012
Blocks:
TreeView+ depends on / blocked
 
Reported: 2019-01-25 07:06 UTC by Rachael
Modified: 2020-02-19 11:19 UTC (History)
15 users (show)

Fixed In Version: oc v3.11.95
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
: 1669560 (view as bug list)
Environment:
Last Closed: 2020-02-19 11:19:52 UTC
Embargoed:


Attachments (Terms of Use)

Comment 6 Niels de Vos 2019-01-25 11:14:27 UTC
(In reply to Xiubo Li from comment #4)
...
> read
> /sys/class/iscsi_host/host415/device/session383/connection383:0/
> iscsi_connection/connection383:0/address: transport endpoint is not connected
> [...]
...

I do not have much experience with the kubernetes iscsi code. But it seems that the above error is not handled correctly at https://github.com/kubernetes/kubernetes/blob/master/pkg/volume/util/device_util_linux.go#L163

From my brief reading of the code, it should not return an error when reading the 'address' file (or others) fail, but instead continue with the loop and look at the next connection.

Humble, what do you think?

Comment 19 RamaKasturi 2019-03-28 14:09:52 UTC
Acking this for 3.11.3


Note You need to log in before you can comment on or make changes to this bug.