In order to have ObjectStore 5.1 manage 2Gb files in a Solaris 2.5.1 environment we have created a 20g striped/mirrored RAW Volume with Veritas not laying a filesystems down on it. The database can be used to organize data on volumes.
---------------------------------------------------------
Procedure: Add 20g raw volume called rawvol1 - striped and mirrored within 'data1' datagroup.
1. highlight c2t8d0s2 and c2t9d0s2 Basic-Ops-->Volume Ops-->Create Striped - " vxassist -g data1 -U fsgen make rawvol1 20g layout=strip,nolog nstripe=2 stripeunit=128 disk2-08 disk2-09 ".
2. hightlight rawvol1 and c3t0d0s2 and c2t1d0s2; Volume Ops-->Add Mirror-->Striped Mirror - " vxassist -g data1 mirror rawvol1 layout=stripe nstrip=2 disk3-00 disk3-01 "
---------------------------------------------------------
Restores of the data was successful and the applications working then last night the CRASH, corrupting many existing filesystems, which we are still running fsck on.
So far the only error message retrieved was from BigBrother.
" WARNING: ufs_readir: bad dir, inumber=574080, fs=/opt/ORACLE7 "
working, more to follow...
P@
---------------------------------------------------------
Procedure: Add 20g raw volume called rawvol1 - striped and mirrored within 'data1' datagroup.
1. highlight c2t8d0s2 and c2t9d0s2 Basic-Ops-->Volume Ops-->Create Striped - " vxassist -g data1 -U fsgen make rawvol1 20g layout=strip,nolog nstripe=2 stripeunit=128 disk2-08 disk2-09 ".
2. hightlight rawvol1 and c3t0d0s2 and c2t1d0s2; Volume Ops-->Add Mirror-->Striped Mirror - " vxassist -g data1 mirror rawvol1 layout=stripe nstrip=2 disk3-00 disk3-01 "
---------------------------------------------------------
Restores of the data was successful and the applications working then last night the CRASH, corrupting many existing filesystems, which we are still running fsck on.
So far the only error message retrieved was from BigBrother.
" WARNING: ufs_readir: bad dir, inumber=574080, fs=/opt/ORACLE7 "
working, more to follow...
P@