I have 1 esx 3.0.1 server connected to MSA 1000. This server uses 2 LUNs on the SAN. Recently I added another esx 3.0.1 server and would like it to use the same 2 LUNS. For some reason I was able to see onle one of the SAN LUNs in the Configuration/Storage window while I see both in Configuration/Storage Adapters pane. Below is the output from esxcfg-mpath -l (from the 'bad' server)
\----
RAID Controller (SCSI-3) vmhba0:0:0 (0MB) has 1 paths and policy of Most Recently Used
FC 6:1.0 210000e08b189f42<->500805f3000db151 vmhba0:0:0 On active preferred
Disk vmhba0:0:3 /dev/sda (140010MB) has 1 paths and policy of Most Recently Used
FC 6:1.0 210000e08b189f42<->500805f3000db151 vmhba0:0:3 On active preferred
Disk vmhba0:0:4 /dev/sdb (420041MB) has 1 paths and policy of Most Recently Used
FC 6:1.0 210000e08b189f42<->500805f3000db151 vmhba0:0:4 On active preferred
Disk vmhba1:0:0 /dev/cciss/c0d0 (69970MB) has 1 paths and policy of Fixed
Local 11:1.0 vmhba1:0:0 On active preferred
\----
The problem is with the Disk vmhba0:0:4 (420041 MB).
Could you please help me to make it visible?
Thank you!
SO, you could leave it this way, but is is not recommended to do so.
This is basically saying that when this LUN was formatted and a signature written on it, it was presented as LUN X, now this LUN is appearing as LUN Y.
I don't know about the MSA 1000, but in the EVA (HP SANs) in the management GUI, when I present a LUN to a HOSt, I can choose which LUN I would like it to be presented as to that host. I alway make sure that all of the hosts that I present this LUN to have the same LUN.
Anyone know how to do this for the MSA 1000??
Is there anything in the logs?
Which logs should I check? (I am the beginner:)
Thank you!
In the VC client go to Task & Events and look for any events listed
Disk vmhba0:0:3 /dev/sda (140010MB) has 1 paths and
policy of Most Recently Used
FC 6:1.0 210000e08b189f42<->500805f3000db151
vmhba0:0:3 On active preferred
Disk vmhba0:0:4 /dev/sdb (420041MB) has 1 paths and
policy of Most Recently Used
FC 6:1.0 210000e08b189f42<->500805f3000db151
vmhba0:0:4 On active preferred
Does it look similar on the other host (especially vmhba0:0:3 and vmhba0:0:4) ?
Was the MSA setup with SSP and the connection type set to Linux?
warning:LVM:4903:vmhba0:0:4:1 may be a snapshot: disabling access. See resignaturing section in SAN config guide. (0:16:20:46.234 cpu2:1037)
Thank you!
Yes.
I would say so. The SAN Identifier is different but everything else is the same.
Below is the output from the 'good' server:
\----
RAID Controller (SCSI-3) vmhba0:0:0 (0MB) has 1 paths and policy of Most Recently Used
FC 6:1.0 210000e08b180a45<->500805f3000db151 vmhba0:0:0 On active preferred
Disk vmhba0:0:3 /dev/sda (140010MB) has 1 paths and policy of Most Recently Used
FC 6:1.0 210000e08b180a45<->500805f3000db151 vmhba0:0:3 On active preferred
Disk vmhba0:0:4 /dev/sdb (420041MB) has 1 paths and policy of Most Recently Used
FC 6:1.0 210000e08b180a45<->500805f3000db151 vmhba0:0:4 On active preferred
Disk vmhba1:0:0 /dev/cciss/c0d0 (69970MB) has 1 paths and policy of Fixed
Local 11:1.0 vmhba1:0:0 On active preferred
\----
Thanks!
warning:LVM:4903:vmhba0:0:4:1 may be a snapshot:
disabling access. See resignaturing section in SAN
config guide. (0:16:20:46.234 cpu2:1037)
Such thinks happens often when the lun is presented to the hosts with different lun numbers - are they equal there ?
On both servers in 'Configuration/Storage adapters' I see the LUN ID 4 for this 'invisible' in Configuration/Storage LUN.
THe "..resignaturing.. & .. snapshot LUN.." message, makes me think that there is a LUN # mismatch. To test this, go to the Configuration Tab in the Infrastructure Client, go to Advanced Settings, click on LVM, and change LVM.DisallowSnapshotLun from 1 to 0.
If after changing to 0, you can see the LUN, then there's a definite mismatch in LUN number from when the LUN was first created to now.
If possible, I would unpresent the LUN from both Hosts (through the SAN management interface) and re-present the LUN making sure to use the same LUN for both of the hosts.
Mayur
THe "..resignaturing.. & .. snapshot LUN.." message,
makes me think that there is a LUN # mismatch. To
test this, go to the Configuration Tab in the
Infrastructure Client, go to Advanced Settings, click
on LVM, and change LVM.DisallowSnapshotLun from 1 to
0.
If after changing to 0, you can see the LUN, then
there's a definite mismatch in LUN number from when
the LUN was first created to now.
If possible, I would unpresent the LUN from both
Hosts (through the SAN management interface) and
re-present the LUN making sure to use the same LUN
for both of the hosts.
Mayur
You are right! I changed LVM.DisallowSnapshotLun from 1 to
0 and I was able to see the second LUN. Could I change the value back? Will I lose it again? Could I leave everything as it is? How will it affect functionality?
If I do unpresent-re-present operation how can I be sure that I am using the same LUN on both hosts? I used SSP on MSA 1000 for configuration.
Thank you!
SO, you could leave it this way, but is is not recommended to do so.
This is basically saying that when this LUN was formatted and a signature written on it, it was presented as LUN X, now this LUN is appearing as LUN Y.
I don't know about the MSA 1000, but in the EVA (HP SANs) in the management GUI, when I present a LUN to a HOSt, I can choose which LUN I would like it to be presented as to that host. I alway make sure that all of the hosts that I present this LUN to have the same LUN.
Anyone know how to do this for the MSA 1000??
Change the setting back now that it is working
I opened case with VMWare techsupport. Specialist provided me with KB6482648 and KB9453805 explaining how to resignature all VMFS volumes, however it is not what I planned to hear. We could not figure out why invisible LUN was detected as snapshot by ESX. I am using the simplest MSA1000 configuration with one switch, no storage groups, one HBA in each of 4 servers ( Exchange, ArcServe Backup, two ESXs).
I have to schedule downtime for all my VMs for resignaturing sometimes next week so if someone knows better solution please let me know! Thanks!
Change the setting back now that it is working
I changed LVM.DisallowSnapshotLun setting back to 1, refreshed the storage and my LUN again disappeared.
I am having the same issue. When I change it back, the LUN is gone. (This is on the local esx box's storage.
I fixed the issue with my local storage, Here is what I did in VC.
1. Unplug the fiber from the SAN. So, you only mess with the local storage.
2. Go to esxc in VC, Click on "Configuration", "Advanced", "LVM" and Change:
A. "LVM.DisallowSnapshotLun" from 1 to 0 (This will allow you to see the LUN)
B. "LVM.EnableResignature" from 0 to 1. (When esx is rebooted, the datastore was named "ESX Local Storage" was changed to "snap000000x-ESX Local Storage" after a refresh. You have to manually change the name back.
3. Once you reboot and can now see the LUN, change A and B above back to the correct inital values. Reboot.
4. ESX will now see the LUNS correctly.
