I know it's not supported but we have an older CX3-80 with NaviSphere 2.24, which has the same HP-UX server in 2 x seperate Storage Groups. The Server sees BOTH sets of LUNs from both Storage Groups. I believe that this is NOT supported, is "illegal", and shouldn't even work, but it's there and working none-the-less.
So, I want to fix this, before we upgrade to FLARE 3.29 (whatever the latest supported is), which we've scheduled for next month. I'm not sure that the EMC CE will do the FLARE Upgrade if the storage configuration is "illegal". They run a HealthCheck, and the config may "fail" the HealthCheck.
Here is what I'm thinking I'll do:
1. Shut down server applications.
2. I'm going to move all of the LUNs from SG2 to SG1 and add them to the existing LUNs in SG1.
3. Then I'm going to destroy SG2.
remove host from SG2.
remove LUNs from SG2
destroy SG2
4. Then, rediscover all LUNs on Server:
ioscan
insf
powermt check
powermt restore
powermt config
powermt save
So, here are some questions:
1. How did the state above come about? It should be impossible.
2. Do you think that the CE's HealthCheck will fail? I'm not sure how smart it is....
2. does this sound like a workable plan, above.
3. will the CxTxDx numbers change on the "moved" SG2 LUNs when I rediscover them? I think that it will, since the "HLU" numbers will change.
4. Will I need to do a "vgexport / vgimport" on the VGs after I rediscover the LUNs on the Server (if the CxTxDx numbers change)?
is it really the same host in two storage groups ? I don't see how that's possible, it should have gotten removed from another SG. Can you check something ..go into SG properties, HOsts tab ..switch to engineering mode (Cntrl + Shift + F12). Compare WWN for host in SG1 and SG2 ..are they really the same ?
I cant answer the HP-UX stuff but you can have a host in 2 different SG's. When you go to add that host to the second storage group, you wont be able to see it till you select from the drop down menu "Hosts already connected" or something like that. Might be all hosts. That window on the left only shows you hosts that are not in a SG.
Soem people use the 2 SG for hosts that are boot from SAN. they make a first SG for the Boot OS lun. then a second for the data lun's
So a host in multiple SG is not bad but the same lun in multiple SG's is bad.
1. How did the state above come about? It should be impossible.
-I've seen this happen when the host is manually registered. For example, each HBA gets registered as a seperate hosts. Not supported but I've seen it.
each HBA to a separate host sure ..but not for the same host.
1. How did the state above come about? It should be impossible.
-I've seen this happen when the host is manually registered. For example, each HBA gets registered as a seperate hosts. Not supported but I've seen it.
2. Do you think that the CE's HealthCheck will fail? I'm not sure how smart it is....
- Probably won't catch this.You can only upgrade to R26 not R29 with a CX3
2. does this sound like a workable plan, above.
-Your plan seems good
3. will the CxTxDx numbers change on the "moved" SG2 LUNs when I rediscover them? I think that it will, since the "HLU" numbers will change.
-Yes
4. Will I need to do a "vgexport / vgimport" on the VGs after I rediscover the LUNs on the Server (if the CxTxDx numbers change)?
-definitley export and import. It will save a lot of grief. Also check for NO_HW devices after the move and removed them
dynamox
11 Legend
•
20419 Posts
•
87439 Points
545
0
Posted March 31st, 2011 07:00
is it really the same host in two storage groups ? I don't see how that's possible, it should have gotten removed from another SG. Can you check something ..go into SG properties, HOsts tab ..switch to engineering mode (Cntrl + Shift + F12). Compare WWN for host in SG1 and SG2 ..are they really the same ?