This post is more than 5 years old
207 Posts
0
667
March 31st, 2011 16:00
VCM disk?
I have a new AIX 6x server connecting to a dmx4. We have not installed Power Path yet. I have mapped and masked 39 8-way striped meta volumes to the HBAs and storage ports.
When I run lsdev -Cc disk it shows me 40 FC disks. I beleive one is a very small disk - maybe a VCM disk. It is showing up as hdisk0. If I remove this disk with rmdev -Rdl hdisk0, it will show up again on the next reboot or cfgmgr. Can I get rid of this disk? I never mapped or masked this disk to the host.
No Events found!
dynamox
9 Legend
•
20.4K Posts
0
April 5th, 2011 14:00
it's the VCM device, when your TC creates a bin file they like to map VCM to all FAs on the array. This device does not have to be masked to appear on the host, any host zoned to that FA will see this device. The only way to get rid of it is to unmap it from the FA. Like i mentioned above if this VCM device has already been discovered by other systems, it may require some clean up before/after you unmap this device from the FA.
PedalHarder
3 Apprentice
•
465 Posts
0
March 31st, 2011 17:00
You may uunap the VCM device, provided it remains mapped to the FA ports used by the SAN management servers.
dynamox
9 Legend
•
20.4K Posts
0
March 31st, 2011 18:00
be careful with unmapping VCM from FAs that has other systems connected to it, all systems will lose access to VCM. Some are oblivious to it (windows) and on some it may cause devices to get renamed (Linux, AIX).
brad12341
207 Posts
0
April 5th, 2011 13:00
I'm still confused as to what is going on. I have not been able to install EMC ODMs, power path or Solutions Enabler yet until our Security group signs off. Those tools will help me debug once I can install them.
I have a second AIX 6.1 host with the same issue. Both these new AIX 6 hosts are LPARS. We are using VIO Servers and the HBAs run to MDS switches using Virtualization (NPIV). I have mapped and masked two 8-way metas to the new host. For some reason lsdev -Cc disk shows 3 disks rather than 2.
# lsdev -Cc disk
hdisk0 Available 01-T1-01 MPIO Other FC SCSI Disk Drive
hdisk1 Available 01-T1-01 MPIO Other FC SCSI Disk Drive
hdisk2 Available 01-T1-01 MPIO Other FC SCSI Disk Drive
If I run bootinfo -s one each hdisk I can see that hdisk0 is the small 2mb volume (maybe gatekeeper or VCM volume). It shows that hdisk1 and hdisk2 are my 8-way striped metas (69712MB or 68GB).
# bootinfo -s hdisk0
2
# bootinfo -s hdisk1
69712
# bootinfo -s hdisk2
69712
I have no idea where this 2MB hdisk0 came from. I did map/mask hdisk1 and hdisk2. But I did not map or mask the small hdisk0. I don't know if this was part of the original EMC bin file? But no one would have known the wwns of the new HBAs to mask too. If I look at the masking database I can not see any reference to this small volume. So I have no way of umasking it as it is not masked. Any ideas?
brad12341
207 Posts
0
April 5th, 2011 14:00
ok, this is the bit of info I needed to tie this all together. Now that I know this is definetely the VCM disk and was mapped by the TC, I should be able to explain this to the UNIX guys so that they are not alarmed by -"what the heck is this little volume". They have not seen it before because most of our AIX boxes are connected to Clariions.