Unsolved
This post is more than 5 years old
21 Posts
0
2881
December 20th, 2018 06:00
[SCv3000] Fail to re-map a volume (failed to create object)
Hello,
We just bought a SCv3000 Compellent SAN and I've created some volumes.
Our Redhat 7.5 servers are connected through dedicated iSCSI VLANs.
The first 3 volumes I created were happily mapped to *two* servers (this is needed by RHEV) and it's working.
Some other 7 volumes are to be mapped to 2 others servers.
The first server can map to these 7 volumes.
But when trying to map these volumes to the second server, some are working and some are failing.
I noticed that on each volume one can see the active controller (top or bottom).
The volume that can be mapped are only the ones on the top controller.
When trying to use the volumes managed by the bottom controller, I get the error "could not create object, please see logs", and in the logs, it just reads 'Bad LUN'.
On the servers list, I indeed see in the connectivity tab the following :
- iSCSI Fault domain 1 : two correct links on the top controller
- iSCSI Fault domain 2 : two correct links on the top controller
I see nothing about the bottom controller (though it is correctly wired).
And I read the message "Only 4 of 8 paths to the server are up".
I see no red cross or nothing else.
The server HBAs list is showing only one HBA of iSCSI port type. (is this normal ?)
On the Linux command, I ran an iscsiadm discovery towards both SAN's targets, and manually ran the connections to the 8 iqn.
Four of them were missing, thus they connected nicely.
Then, on the connectivity tab, the links to the bottom controller appeared.
I was then able to map the missing volumes to this server.
Is this normal to see only one HBA (each Linux server has two HBAs with a dedicated ip on its dedicated VLAN)?
DELL-Bob Mi
2 Intern
•
230 Posts
0
December 22nd, 2018 10:00
First I would look at the SCv3000 and SCv3020 Storage System Deployment Guide and verify your iSCSI setup versus what is listed in iSCSI Settings Appendix E. Verify your cabling is done properly, and your settings are correct.
You do not mention MPIO configuration which of course is a necessity. Verify it is configured correctly based upon section 3.6 of the Dell Storage Center with Red Hat Enterprise Linux (RHEL) 7x Best Practices
http://en.community.dell.com/techcenter/extras/m/white_papers/20440304
Verify your server configuration settings in Section 3. You do have to scan for newly mapped volumes on your RHEL server. This is explained in section 2.1 Scanning for new volumes.