Concurrent iSCSI Target Portal Login fails with Cinder HPE 3PAR iSCSI driver
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Cinder |
New
|
Undecided
|
Unassigned |
Bug Description
When multiople cinder volumes are created concurrently from the same source image for the first time (image-
Affected Version: stable/mitaka
Steps to Reproduce:
1. Configure cinder.conf to use HPE3PARISCSIDriver as given in config-reference.
2. chap auth should be enabled and image volume cache should be enabled.
2. Complete all other cinder.conf settings.
[HPE-
san_
hp3par_
volume_
san_
image_
hp3par_
volume_
hp3par_
hp3par_
hp3par_
hp3par_
san_ip=<san_ip>
3. Start all cinder services and ensure the driver gets initialized successfully.
4. Execute the following command from cinder node via CLI:
$ export TYP=<your_vol_type>
$ export IMG=<glance_
$ for i in `seq 15`; do cinder create --image-id $IMG --name $TYP$i --volume-type $TYP 100 & done
5. Execute cinder list and observe the status of the created volumes.
6. Rinse and repeat step 4 and 5 a few times, each time with a different image (to exercise vol-cache-miss flow)
NOTE:
* This issue is NOT seen in HPE3PARFCDriver
* This issue is not reproducible with vol-cache-enabled set to FALSE
* This issue is not seen when exercising vol-cache-hit code path (when image is available in vol backed cache)
description: | updated |