2
Any two or more drives in the array are faulted. Typically this represents from 2 to 6 faulted drives, and not entire enclosures, which would be a different tree. This tree assumes that the disks are still faulted. (0) Yes Are there only 2 drives faulted in the array? (2) No Arrays with Multiple Problems: See Knowledgebase article emc81176 when troubleshooting Fibre Channel arrays with multiple problems. This solution examines the order that the problems should be addressed to maximize the chance of success. Run SPcollect on both SP’s. Analyze information using available tools (Triage, SPLAT, CAP, RLS Collector, etc.). (1) Add check for the case of two drives actually faulted (check for 801/803 errors from those drives only) (4) Are the logs free of: * backend loop errors on other drives, * anything that indicates a problem elsewhere on the loop? (6) Yes Review the spcollect command output from both SPs, and determine which side (A or B) faulted the drives first. Sequence of replacement (continue until drives are no longer faulted and Soft SCSI errors are eliminated): 1. If more than two non-contiguous drives in the same DAE-2P are faulted (possibly along with an LCC), check Primus case emc132617 for a possible power branch problem. 2. Replace the LCC on the side that the drives were first faulted. If it is an ATA enclosure, replace both BCCs. Reference the CPG for these replacements. Confirm that all drives in this enclosure are no longer faulted, and that this enclosure generates no new Soft SCSI errors. 3. If there are still drives faulted or this enclosure still generates Soft SCSI errors, replace all cables on this side of the bus (if Bus 0, this will cause an SP reboot). 4. If this enclosure still has faulted drives or generates Soft SCSI errors, escalate. (7) Are all faulted drives in the same enclosure? (3) Yes A No B No Do the logs show that these drives faulted at different times? (5) Yes No

EMC clariion multiple_drives_faulted Toubleshooting

  • Upload
    joe-cap

  • View
    42

  • Download
    4

Embed Size (px)

DESCRIPTION

EMC clariion , faults, disk, array

Citation preview

Page 1: EMC clariion multiple_drives_faulted Toubleshooting

Any two or more drives in the array are faulted. Typically this represents from 2 to 6 faulted drives, and not entire enclosures, which would be a

different tree. This tree assumes that the disks are still faulted. (0)

YesAre there only 2 drivesfaulted in the array? (2)

No

Arrays with Multiple Problems: See Knowledgebase article emc81176 when troubleshooting Fibre Channel arrays with multiple problems. This solution examines the order that the problems should be addressed to maximize the chance of success.

Run SPcollect on both SP’s. Analyze information using available tools (Triage, SPLAT, CAP, RLS Collector, etc.). (1)

Add check for the case of two drives actually faulted (check for 801/803 errors from those drives only) (4)

Are the logs free of:* backend loop errors on other drives,

* anything that indicates a problem elsewhere on the loop? (6)

Yes

Review the spcollect command output from both SPs, and determine which side (A or B) faulted the drives first.

Sequence of replacement (continue until drives are no longer faulted and Soft SCSI errors are eliminated):

1. If more than two non-contiguous drives in the same DAE-2P are faulted (possibly along with an LCC), check Primus case emc132617 for a possible power branch problem.

2. Replace the LCC on the side that the drives were first faulted. If it is an ATA enclosure, replace both BCCs. Reference the CPG for these replacements. Confirm that all drives in this enclosure are no longer faulted, and that this enclosure generates no new Soft SCSI errors.

3. If there are still drives faulted or this enclosure still generates Soft SCSI errors, replace all cables on this side of the bus (if Bus 0, this will cause an SP reboot).

4. If this enclosure still has faulted drives or generates Soft SCSI errors, escalate. (7)

Are all faulted drives inthe same enclosure? (3)

Yes

A No

B

No

Do the logs show that these drives faulted at different times? (5)

Yes

No

Page 2: EMC clariion multiple_drives_faulted Toubleshooting

Yes

Are all faulted drives in different enclosures on the

same BE Bus? (8)No

A

There are multiple drive failures on multiple BE buses.Escalate to the appropriate level. (9)

NOTE: When closing a solved case, be sure to open Clarify and specify the action that solved the problem. If the case needs to be escalated, make sure that you have the spcollect data.

Are the drives in the same raid group? (11)

This may be a 2-drive failure.Recovery procedure may be available.Escalate to the appropriate level. (13)

YesReplace both drives.

Refer to the Single Drive Failuretroubleshooting tree. (12)

No

B

Review spcollect output from both SPs, and determine which side (A or B) faulted drives first.

Sequence of replacement (continue until drives are no longer faulted and Soft SCSI errors are eliminated): 1. Replace all cables on this side of the bus (if Bus 0, this will cause an SP reboot). 2. Escalate to the appropriate level. Order enough FRUs to replace all LCCs/BCCs on the side that the drives were first faulted (order BCCs for both sides). Reference CPG for these replacements. (10)