Unsolved
This post is more than 5 years old
38 Posts
0
652
Bypass DAE2 on SPA only - diag soft SCSI errors
Hi All,
Can a DAE2 be bypassed on the SPA side while leaving side SPB active?
I have a CX 500 that is suffering from excessive soft scsi bus errors on Bus 0 enclosures, but only when the LUNs on those enclosures are on SPA.
Bus 0
DPE
DAE2
DAE2P
Bus 1
DPE
DAE2P
If I move a LUN(s) on a Bus 0 enclosure to SPB, the soft scsi bus errors go away for that LUN.
SPA has been replaced
LLC cards have been replaced on SPA side of enclosures
cables have been replaced on SPA side of enclosures
Dell/EMC believe that the problem may be the DAE2 enclosure at Bus 0, enclosure 1. Since replcement of that enclosure will require downtime, I'd like to be doubly-sure that the problem resides with that enclosure.
If I were to trespass everything on Bus 0 (all enclosures) from SPA to SPB, could I then temporarilly disconnect (cable around) Bus 0 Enclosure 1 on the SPA side while leaving it connected on SPB? Once bypassed, I want to trespass Bus 0 enclosure 0, and Bus 0 enclosure 2 LUNs back to SPA and then test for soft scsi bus errors. The idea being that if the soft scsi errors stop for those enclosures, then I'm more confident of the issue being with Bus 0 Enclosure 1.
Once the testing is done, I'd trespass everything back to SPB, then re-add bus 0 enclosure 1 to the SPA side.
thoughts?
Message was edited by: JeffScripps
nandas
1.5K Posts
0
December 14th, 2009 10:00
Hi Jeff,
Yes - the problem may be caused by one side of the connection namely SPA side of connection here. The cable, LCC, Firmware on LCC, Enclosure backplane etc. may come in to the picture for this type of issue.
Are you seeing the errors on the LUNs on the DAE2P enclosure only? If that is the case, the next logical step will be to replace the enclosure. However, it is very difficult to comment like this without seeing any details or logs etc. I am sure the Dell/EMC support team will guide you to the right direction.
Regarding your testing of bypassing SPA side of connectivity on Bus 0 - this is little bit risky with any live (production) data, and not recommended in a production environment. However, the LUNs should be available through SPB if they are owned by SPB and SPB side of connectivity is intact. Now, keep in mind that this is CX500 (old hardware and old Bus technology). If you bypass the SPA side of BUS0 connection on Enclosure 1 - automatically the SPA side of connection for Enclosure 2 is also disconnected. So, you can not trespass back the LUNs from SPB to SPB which resides on enclosure 1 or enclosure 2. You may think of disconnecting only Enclosure 2 SPA side connection and keep Enclosure 1 SPA side connection intact. Then trespass the LUNs on Enclosure 1 to SPA and check. However, I 'll suggest please discuss your ideas or plans with Dell/EMC team who are working on this issue before doing anything further.
Thanks,
Sandip