S2600 new disk can't access to system

Publication Date:  2013-01-24 Views:  160 Downloads:  0
Issue Description
One S2600 device has a failed disk, engineer applied for a new disk to replace it , but could not be identified by system
Alarm Information
no
Handling Process
through the analysis to the logs , we got the following alarms when disk access in.

Jan  8 11:44:18 AK-I kernel: [31711070353]mptbase: ioc0: WARNING - Called 3 event handlers, sum results = 2
Jan  8 11:44:18 AK-I kernel: [31711070353]mptbase: ioc0: WARNING - EVENT_NOTIFICATION reply 0000010032800ff0 freereq=0
Jan  8 11:44:18 AK-I kernel: [31711070356]INI_JamReqScmd:663: INFO: ioc0 cmd ffffff00004cf678(96:101003102c8:0)(0:6) to disk 2:0:105 is jammed (0:0)
Jan  8 11:44:18 AK-I kernel: [31711070362]mptbase: ioc0: WARNING - MPT event:(16h) : SAS Discovery: Stop

we can conclude it was caued by  controller starter chip abnormal of S2600. Have to upgrade the system firmware version to solve this problem, or restart starter chip to avoid it temporary.

how to upgrade :
1. plug out the disk that can't access to system normally, stop the bussiness;
2. reboot the system to make the initiator to be normal
3. upgrade the version to 1.04.01.212.T08(V100R001C02SPC019)or more higher version with ISM or root script
4. upgrade initiator version to 010 with implementing CLI command upgradeinifw, then implement command showsasmode to check the initiator verion
5. by the way implement the command chgbstswitch -o  to open the BST function.
6. plugin  the disk.

Root Cause
First, suppose it was caused by slot or disk problem;
1. Plug a hotspare disk into this slot belong to this device, it works well, so it was not caused by slot fault.
2. Plug the new spare disk into the solt where the hotspare disk been plugged out, it can't work, so we suppose it maybe caused by the new disk , we applied for another spare disk to try again, it didn't work too, eliminate disk problem.
3. Now, we can suppose it was caused by S2600 system or version problem about disk
4. collect system logs to analyse
Suggestions
no

END