Unsolved
This post is more than 5 years old
22 Posts
0
15621
June 1st, 2010 09:00
Solaris 10 and EMC CX3-40c storage array
When issuing commands in the Soaris O/S such as "format", or updating the "Host Side Information" when in Navisphere Manager, the Solaris console shows drives as being in the "offline state". An example of the error is as follow:
WARNING: /pci@9,600000/SUNW,qlc@1/fp@0,0/ssd@w5006016941e09fea,0 (ssd17): drive offline.
It looks like the system is seeing the Storage Processors (SP) themselves as disk drives. I believe that there is a way to fix this, but I can't determine what needs to be altered so this error will no longer show. What is the resolution for this error?
Does anybody else have or seen this error and know how to resolve these error messages.
The server is a Sun Sunfire V890, running Solaris 10 u5 and Oracle RDBMS 11G R1. The HBA's installed in the server are re-branded Qlogic QLA2460-S HBA's, i.e. SG-XPCI1FC-QF4, connected to CX3-40c via 2 Brocade 5000 FC switches.
Thank you,
Eric Buttrum
Dev Mgr
4 Operator
•
9.3K Posts
0
June 1st, 2010 11:00
Assuming you registered the server correctly, put it in a storage group with at least 1 LUN, I remember that some of the Unix flavors will not keep FC links alive if no disks are mounted, but I don't remember which Unix flavor that was, or what the way is to get the link up to be able to mount a disk/partition.
ewhb
22 Posts
0
June 1st, 2010 12:00
Sorry, I should've been more thorough in my problem description. There are 4 LUNS mounted to the Solaris system and they are seen by Solaris. I have partitioned these 4 drives, created file systems on them and mounted them in Solaris and are in use for Oracle databases. What is happening is I also see 8 warnings that drives are in the "offline" state and 8 "drive type unknown" errors, along with the 16 drives that powerpath sees and he 4 pseudo drives that powerpath creates when running the "format" command.
Thanks,
Eric.
ewhb
22 Posts
0
June 6th, 2010 12:00
I finally solved this issue. This happens because there is no "LUN 0" being associated with the Solaris server from the SAN. One of the LUNS has to be " LUN 0" or these errors show up. I removed the LUNS from the Solaris server and performed a "boot -r" and issued "powermt" commands to remove all remnants of the SAN LUNS. Then I associated the LUNS again with one of the LUNS set as LUN 0 and then performed a "boot -r" again and then issued the "powermt" commands again. The LUNS are seen again with the same pseudo "emcpower" names and the errors of drives being offline are no longer present.