i all,
I want to put a local disk (DGLOCAL) on a Sun Cluster node but scconf
command explodes
My system:
* two node cluster on two VMWare virtual machines
* Solaris 10 SunOS 5.10 Generic_141415-05 i86pc i386 i86pc
* Sun Cluster 3.2 u2
* Veritas Volume manager
The situation is:
----------------------------------------------------
# vxdisk list
DEVICE TYPE DISK GROUP STATUS
c1t0d0s2 auto:none - - online invalid
c1t1d0s2 auto:none - - online invalid
c2t0d0s2 auto:cdsdisk DGLOCAL01 DGLOCAL online
c2t1d0s2 auto:cdsdisk DGCLUSTER01 DGCLUSTER online
...cut...
# vxdisk -g DGLOCAL check DGLOCAL01
c2t0d0s2: Okay
# didadm -l
8 clab1:/dev/rdsk/c2t0d0 /dev/did/rdsk/d8
...cut...
# cldev status
Device Instance Node Status
--------------- ---- ------
/dev/did/rdsk/d8 clab1 Ok
...cut...
# fstyp /dev/vx/rdsk/DGLOCAL/DGLOCAL_VOL1
ufs
----------------------------------------------------
Test mount ok.
Is not an hard disk problem.
When I try:
# scconf -a -D type=vxvm,name=DGLOCAL,localonly=true
Segmentation Fault (core dumped)
truss output:
----------------------------------------------------
...cut...
xstat(2, "/opt/VRTSvxvm/lib/libvxvmsc.so", 0x080462C0) = 0
resolvepath("/opt/VRTSvxvm/lib/libvxvmsc.so", "/opt/VRTSvxvm/lib/
libvxvmsc.so", 1023) = 30
open("/opt/VRTSvxvm/lib/libvxvmsc.so", O_RDONLY) = 5
mmap(0x00010000, 32768, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_ALIGN, 5, 0)
= 0xCE220000
mmap(0x00010000, 200704, PROT_NONE, MAP_PRIVATE|MAP_NORESERVE|MAP_ANON|
MAP_ALIGN, -1, 0) = 0xCDC30000
mmap(0xCDC30000, 125325, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_FIXED|
MAP_TEXT, 5, 0) = 0xCDC30000
mmap(0xCDC5E000, 10982, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE|
MAP_FIXED|MAP_INITDATA, 5, 122880) = 0xCDC5E000
munmap(0xCDC4F000, 61440) = 0
memcntl(0xCDC30000, 15708, MC_ADVISE, MADV_WILLNEED, 0, 0) = 0
close(5) = 0
munmap(0xCE220000, 32768) = 0
Incurred fault #6, FLTBOUNDS %pc = 0xCDC91FA9
siginfo: SIGSEGV SEGV_MAPERR addr=0x00000000
Received signal #11, SIGSEGV [default]
siginfo: SIGSEGV SEGV_MAPERR addr=0x00000000
----------------------------------------------------
core output:
----------------------------------------------------
core 'core' of 6256: scconf -a -D type=vxvm,name=DGLOCAL,localonly=true
cdc91fa9 scds_add_ds_local (809bae7, 0, 2, 0, 2, 0) + 451
cdc92711 scds_add_ds (809bae7, 0, 0, 0, 2, 0) + 135
ceb93619 scconf_add_ds_vers2 (809badd, 809bae7, 0, 0, 0, 2) + 79
0805ca3b ???????? (8047e3d, 8047d40, 0)
08053235 ???????? (4, 8047d7c, 8047d40, 0)
08052853 main (4, 8047d7c, 8047d90) + 363
0805245a ???????? (4, 8047e30, 8047e37, 8047e3a, 8047e3d, 0)
----------------------------------------------------
/var/adm/messages silent
I have also tried to change the Veritas disk naming scheme from enclosure
based OFF to ON (although this is not a problem with WWN names).
PS: shared disk (DGCLUSTER) work fine
# cldg status
=== Cluster Device Groups ===
--- Device Group Status ---
Device Group Name Primary Secondary Status
----------------- ------- --------- ------
DGCLUSTER clab1 c