Bug ID: 6561826 mirrors in a metaset may not be resynced on a take of the set

Bug ID	6561826
Synopsis	mirrors in a metaset may not be resynced on a take of the set
State	10-Fix Delivered (Fix available in build)
Category:Subcategory	kernel:svm
Keywords	rtiq_reviewed \| scnv
Responsible Engineer	Peter Dennis
Reported Against	3.2_fcs
Duplicate Of
Introduced In	solaris_10
Commit to Fix	snv_68
Fixed In	snv_68
Release Fixed	solaris_nevada(snv_68) , solaris_10u7(s10u7_01) (Bug ID:2150760)
Related Bugs	4518073 , 6627803
Submit Date	24-May-2007
Last Update Date	11-July-2007
Description	Please login to pkuda1 to take a look.. This is a test stopper for OSC testing. metaset kuda-2 is working fine. But, sets kuda-1 and kuda-3 are having this problem. # /usr/cluster/bin/scstat -D -- Device Group Servers -- Device Group Primary Secondary ------------ ------- --------- Device group servers: kuda-1 pkuda1 pkuda2 Device group servers: kuda-2 pkuda1 pkuda2 Device group servers: kuda-3 pkuda1 pkuda2 -- Device Group Status -- Device Group Status ------------ ------ Device group status: kuda-1 Online Device group status: kuda-2 Online Device group status: kuda-3 Online -- Multi-owner Device Groups -- Device Group Online Status ------------ ------------- # newfs -v /dev/md/kuda-1/rdsk/d0 newfs: construct a new file system /dev/md/kuda-1/rdsk/d0: (y/n)? y mkfs -F ufs /dev/md/kuda-1/rdsk/d0 104824832 256 -1 8192 1024 220 1 120 8192 t 0 -1 8 7 n /dev/md/kuda-1/rdsk/d0: cannot create: Read-only file system # df -kl\|grep global /dev/did/dsk/d17s3 980920 4997 917068 1% /global/.devices/node@4 /dev/did/dsk/d1s3 980920 5009 917056 1% /global/.devices/node@1 /dev/did/dsk/d13s3 980920 5001 917064 1% /global/.devices/node@2 /dev/did/dsk/d15s3 980920 4997 917068 1% /global/.devices/node@3 A switchover after mounting a global FS is causing this I/O error. May 24 09:55:19 pkuda2 last message repeated 21 times May 24 09:55:48 pkuda2 genunix: WARNING: Switchover (/global/test-fs-2) error: failed to mount FS (30) May 24 09:55:50 pkuda2 syseventd[2071]: /devices or /dev may not be current (devfsadmd not responding) Run devfsadm These errors may not be causing this bug, but there are some warnings on the console whenever devfsadm is run. =============================================================================== Slow bootup due probably to these warnings ( bugid : 6550591 ) # devfsadm May 24 09:59:55 pkuda2 scsi: WARNING: /pci@0,0/pci1022,7450@a/pci17c2,20@4 (mpt0): May 24 09:59:55 pkuda2 mpt driver unable to online target 6 lun 0 # echo $? 0 =========================================================================== Clarifying the description here - when a diskset is taken via metaset -t the mirrors that are defined in the set may be left in the 'needs maintenance' state: root@vha-v65xa> metastat -s kuda-2 kuda-2/d0: Mirror Submirror 0: kuda-2/d11 State: Needs maintenance Submirror 1: kuda-2/d12 State: Needs maintenance Pass: 1 Read option: roundrobin (default) Write option: parallel (default) Size: 71115624 blocks (33 GB) kuda-2/d11: Submirror of kuda-2/d0 State: Needs maintenance Invoke: metasync kuda-2/d0 Size: 71115624 blocks (33 GB) Stripe 0: Device Start Block Dbase State Reloc Hot Spare c0t3d0s0 0 No Okay Yes kuda-2/d12: Submirror of kuda-2/d0 State: Needs maintenance Invoke: metasync kuda-2/d0 Size: 71115624 blocks (33 GB) Stripe 0: Device Start Block Dbase State Reloc Hot Spare c0t4d0s0 0 No Okay Yes The reason for this is that the devfsadmd has not created the /dev/md links by the time the resync process has started. Therefore the resync process cannot open the devices to kick off the resyncs.
Work Around	N/A
Comments	N/A

OpenSolaris