Bus shows intermittently failed

If it is interment issue, my suggestion 1) restart SPs management server one by one

                                                         2)  reboot of SPs One by one at time ( before reboot SP trespass LUNs

                                                         3) replace LCC from that encl.

I would recommend opening a case with EMC - bus errors could be coming from either SPA and SPB, cables, LCC's or disks.

glen

Hi All,

We are seeing errors in navisphere manager for this enclosure, this enclosure was added using NST. Error only reported by SP A and not by SP B. After analysis on logs support suggested to reseat LCC A on this enclsoure after reseating LCC A this enclosure is now reported as missing  and LUN's bound from it are in Degraded mode.

Management server was restarted for both SP's but it didnt clear the error, now support suggests power cycle full array which is quite not possible for the customer as applications are 24*7 in production.

Any help or suggestions would be appreciated.

regards,

Samir

Try powering off the new DAE (both A and B side at same time) - as long as it is the last device on the loop. Last thing you can do before rebooting the SP.

glen

Hi all,

One thing I noticed was that while adding this DAE-ATA NST assigned it with Bus 1 Encl 4 but number on it was "2" at the rear of this DAE it was recognized on the array, we configured raid group using this enclosure disks and luns were assigned to host. Now after this error message was seen after LCC-A reseat on this enclosure Bus 1 Encl 4 went missing but there appears Bus 1 Encl 2 (which was not used earlier) with same number of disks, I informed support on this behaviour and asked the engineer if I can power off this DAE assign enclosure id as 4 instead of 2 and lets check if its correctly recognized on the array or not, the engineer said not to try it but instead power cycle the array.

Want to know if trying it out would cause more complications?

regards,

Samir

If i remember right,  changing the enclosure ID after RG's and Lun's have been bound will corrupt the data on them.

The NST looks for the next logical BUS_ENCL from your current config to assign to the next one.  You have to manually set the enclosure ID as one of the steps.

EMC wants you to reboot and when it does, that DAE will be at bus1_encl2.  the only way i can think of changing it would be to move all data off of it and destroy all RG's and luns.   then remove it from san and add it back after you change the enclosure id to 4

did you already have a bus 1 encl 2 & 3?  i would think so since the NST wanted you to add it as 4    That would be bad news if you added this DAE on the same bus as enclosure 2 when there was already one present

Bus 1 Encl 2 was never used when this DAE was added. there were two DAE-ATA's added one was assigned Bus 0 Encl 2 and other was assigned Bus 1 Encl 4 by NST.It all worked for over a week there are lun's bound and being used by host.So even after power cycle of this enclosure DAE will still assign it Bus 1 Encl 2 as the id coz encl is 2 from the enclosure.

So if I and support wants to clear this error:-

1. Move/copy the Lun's to another space using this enclosure.

2. Unbind luns from this enclosure.

3. Destroy this RG.

4. Power off this DAE.

5. Disconnect this DAE from CX700.

6. Power cycle the array.

7. Change Encl id of this DAE to 4 instead of 2(currently set).

8. Use NST to add this DAE.

9. Assign enclosure id as stated by NST and physically change it on DAE.

10. Connect B-side and than A-side BE cable as instructed by NST.

11. Than create RG and Lun's.

12.Assign Lun to host.

Hope this steps are correct, please correct me if I missed out on any step? Also would like to know if I would be able to migrate these LUN's currently running in Degraded mode?

regards,

Samir

The steps look good to me - you want to change the DAE address before you add it back into the array. Even in degradef mode you should be able to copy the data off the LUNs - degraded can mean that one side is still working - may need to trespass the LUNs to the working side - open two Navisphere Windows - one to SPA and the other to SPB - you will get the view of the disks from each SP - it may be missing on one SP but visible on the other.

glen

Hi Glen,

Support suggests not to do any changes until box is rebooted. But I feel this DAE shall still have Bus 1 Encl 2 as the id.

Any suggestions would be appreciated.

regards,

Samir

At this point, you should probably follow support's action plan - not having access to the same data that they do, I would not want to give you incorrect information.

glen

We have scheduled power cycle of CX700 on Friday morning.

shall post the results after it.

regards,
Samir

Hi,

As support suggested power cycled the array > all ghost entries were removed after powerup.

But this enclosure still showed faulted from SPA. I changed enclosure id to 4 and issue got resolved.

Thanks for all your suggestions.

regards,

Samir