Ubuntu: Storage property of deployed MAAS node report 0 disks [on hold]



Question:

My controlnode is a hyper-v vm image running Trusty LTS with the maas maas-region-controller maas-cluster-control and juju that control 4 physical HP ProLiant DL360G5 nodes in a ready state. I successfully commissioned each of them with the debian installer. My default deployment iso is Trusty 14.10. When I go to the machine summary of each node, all information is correct except for the storage details on all 4 nodes, which always shows 0.0GB over 0 disks.

When I ssh'ed into one of these nodes and executed a lshw -class disk command, lshw did not find any disks on the system. However lshw -class storage did find my HP P400i scsi raid controller.

When I cd to /dev/disk/by-path I do see them and the deployed node shows it is installed on disk 0 where I see that it created 3 partitions.

pci-0000:06:00.0-cciss-disk0  pci-0000:06:00.0-cciss-disk0-part1  pci-0000:06:00.0-cciss-disk0-part2  pci-0000:06:00.0-cciss-disk0-part5  pci-0000:06:00.0-cciss-disk1  pci-0000:06:00.0-cciss-disk2  pci-0000:06:00.0-cciss-disk3  pci-0000:06:00.0-cciss-disk4  pci-0000:06:00.0-cciss-disk5  

That made me also realize that the disks are not called /dev/sda, /dev/sdb, etc, but instead /dev/cciss/c0d0, /dev/cciss/c0d1 etc. It seems to me that lshw is the problem in this matter.

I am also curious if it would be possible, in case I find a fixed version of lshw, to manually update the deployment images with another version of lshw and if that can be done with the juju deployment image as well?

Update: In the mean time, I went to the authors website of lshw and it seems that they have problems, so I could not enter a ticket. I could however jump into the source code and found that the struct *devices in http://ezix.org/source/packages/lshw/releases/B.02.17/src/core/scsi.cc does not contain a search query to /dev/cciss/c* which I think now is the problem.

Update 2: Reported bug at https://bugs.launchpad.net/maas/+bug/1476291

Regards, Joham


Solution:1

This problem is fixed in MAAS version 1.8.0+bzr4001-0ubuntu2 (trusty1). All nodes and its storage are identified correctly when I recommissioned the machines after upgrading from MAAS 1.7.


Note:If u also have question or solution just comment us below or mail us on toontricks1994@gmail.com
Previous
Next Post »