DISK PROBLEMS SOLARIS 8

[mahalingam1]

Hi all, I have a sun-fire-480R under solaris 8.

Yesterday we had a disk problem. The disk had a lot of transport errors. This disk is under solaris volume manager

The disk was changed but the system still see it.

$ iostat -Enx
...
...
ssd2 Soft Errors: 7 Hard Errors: 1 Transport Errors: 1260
Vendor: SEAGATE Product: ST336605FSUN36G Revision: 0638 Serial No: 0221P1Q2A6
Size: 36.42GB <36418595328 bytes>
Media Error: 0 Device Not Ready: 0 No Device: 1 Recoverable: 7
Illegal Request: 0 Predictive Failure Analysis: 0

I want know if is possible to do the system not see the disk wihtout reboot the system

I know it's possible in solaris 9 and solaris 10 but i'm don't know how to do it. i'm not sure if it's possible in solaris 8

Please anyone can help me???

Thanks in advance and regards

 

[Mike042]

>> The disk was changed but the system still see it.

It seems you packed a lot in, in the above statement. There is normally quite a long process to swap out a 'failed' disk under SVM (in Solaris 9 & 10) or under SDS (in Solaris 8).
The steps would include:
- deleting the stale State Database Replicas
- detaching the submirrors
- clearing the submirrors
- unconfiguring the failed disk (via cfgadm)
- swap the failed disk for its replacement
- configure the new disk (via cfgadm)
- re-create the State Database Replicas
- create and attach the submirrors

Following this method it is possible the change a 'failed disk', without a reboot. Provided, of course, the 'User Guides' say that the disks are 'hot-swappable'.


I hope that helps.

Mike

 

[mahalingam1]

Hi Mike 42, thanks for your help

I have the following problem whith cfgadm command. This command not recognizes this disk. It's very curious but the disk belonging to metadevices are not list via cfgadm

cfgadm -al
Ap_Id Type Receptacle Occupant Condition
c0 scsi-bus connected configured unknown
c0::dsk/c0t0d0 CD-ROM connected configured unknown
c2 scsi-bus connected unconfigured unknown
c3 scsi-bus connected unconfigured unknown


# format
Searching for disks...done


AVAILABLE DISK SELECTIONS:
0. c1t0d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@9,600000/SUNW,qlc@2/fp@0,0/ssd@w21000018625d2412,0
1. c1t1d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/pci@9,600000/SUNW,qlc@2/fp@0,0/ssd@w21000004cffb7cf5,0



# metastat -p
d10 -m d12 d11 1
d12 1 1 c1t1d0s0
d11 1 1 c1t0d0s0
d20 -m d22 d21 1
d22 1 1 c1t1d0s1
d21 1 1 c1t0d0s1
d30 -m d32 d31 1
d32 1 1 c1t1d0s3
d31 1 1 c1t0d0s3
d40 -m d42 d41 1
d42 1 1 c1t1d0s4
d41 1 1 c1t0d0s4
d50 -m d52 d51 1
d52 1 1 c1t1d0s5
d51 1 1 c1t0d0s5



know what is happening???

Thank you very much for you help

regards

 

[Mike042]

It looks like your two disks are 'fibre-attached'. I have only ever used SDS (Solaris 8) and SVS (Solaris 10) on the internal IDE or SCSI disks. I am puzzled as to why 'c1' is not being reported by cfgadm. Additionally I am not familiar with the SunFire 480R hardware, as to whether 'hot-swapping' disks is allowed. (If it is not, then a reboot may be the only option, after ascertaining that 'c1' scsi-bus is Ok and the devices can be 'seen').


I hope that helps.

Mike

 

[elgrandeperro]

I am pretty sure 480R is FCAL (from the ssd line in the device path) and that it is hot swap. I think you can ignore the cfgadm stuff.

I think a few minor steps are missing:

...
- configure the new disk (via cfgadm)
- CONFIGURE THE SLICES USING PRTVTOC
- INSTALL BOOTBLOCK (if root mirror)
- re-create the State Database Replicas
- create and attach the submirrors
- RUN METASTAT TO CHECK STATUS AND SYNCING

 

[jimmies]

Hi
You should use luxadm command instead of cfgadm and proceed with the same steps described by Mike042.
But unfortunately the iostat command will still report the errors for the old disk and only a reboot will get rid off these errors.
Cheers

 

[mahalingam1]

Hi all, thanks for your information. Finally we have rebooted the server and the problems was solved.

Nevertheless the disk belonging to all metadevice are not recognizes by cfgadm command.


Thanks again and regards