A failure occurred during snapshot creation on an Access 7.4.1.200 cluster and rollback wasn't successful.
Commands used to try and remove the snapshot and get the configuration back into a clean step resulted in stale objects which couldn't be removed. These stale objects were impacting backup and other snapshot activities.
# vxprint -g sfsdg -qth
v SNAP-kavol_tier1 - DISABLED CLEAN 2097152 SELECT - fsgen <<<<<< stale
pl SNAP-kavol_tier1-P01 SNAP-kavol_tier1 DETACHED STALE 2097152 CONCAT - RW <<<<<< stale
sc SNAP-kavol_tier1-S01 SNAP-kavol_tier1-P01 kacache1 0 2097152 0 - DIS <<<<<< stale
vt kavol - ENABLED ACTIVE 1
v kavol_tier1 kavol ENABLED ACTIVE 2097152 SELECT - fsgen
pl kavol_tier1-01 kavol_tier1 ENABLED ACTIVE 2097152 CONCAT - RW
sd emc0_0391-01 kavol_tier1-01 emc0_0391 0 2097152 0 emc0_0391 ENA
dc kavol_tier1_dco kavol_tier1 kavol_tier1_dcl
v kavol_tier1_dcl - ENABLED ACTIVE 67840 SELECT - gen
pl kavol_tier1_dcl-01 kavol_tier1_dcl ENABLED ACTIVE 67840 CONCAT - RW
sd emc0_0394-01 kavol_tier1_dcl-01 emc0_0394 0 67840 0 emc0_0394 ENA
pl kavol_tier1_dcl-02 kavol_tier1_dcl ENABLED ACTIVE 67840 CONCAT - RW
sd emc0_0395-01 kavol_tier1_dcl-02 emc0_0395 0 67840 0 emc0_0395 ENA
co kacache1 - DISABLED ACTIVE <<<<<< stale
Examples of some of the commands to try and remove the above stale objects:
# vxedit -g sfsdg -rf rm kacache1
VxVM vxedit ERROR V-5-1-6321 Cache kacache1 has associated subcaches
# vxsd -g sfsdg -o force dis SNAP-kavol_tier1-S01
VxVM vxsd ERROR V-5-1-10128 Cache object meta-data update error
In order to resolve this issue a new cache volume was created and associated with the stale cache. This allowed the stale cache to be started followed by the stale snapshot volume. They could then both be successfully removed and no stale objects remained.
1. vxassist -g sfsdg make kacachevol 10g
2. vxcache -g sfsdg att kacachevol kacache1
3. vxcache -g sfsdg start kacache1
4. vxvol -g sfsdg -f start SNAP-kavol_tier1
5. vxedit -g sfsdg -rf rm SNAP-kavol_tier1
6. vxedit -g sfsdg -rf rm kacache1