Service Guard - FS Mount hang

[Jonathan Caplette_1]

Hi guys,

I've a 2 nodes cluster running. I shutted down the package on one node, successfully. Now I try to run the same package on the other node, the filesystems are not mounting it is hang.

########### Node "psa1h01": Starting package at Sat Dec 1 07:29:12 EST 2007 ###########
Dec 1 07:29:12 - Node "psa1h01": Activating volume group vg01PA with exclusive option.
Activated volume group in Exclusive Mode.
Volume group "vg01PA" has been successfully changed.
Dec 1 07:29:12 - Node "psa1h01": Activating volume group vg01PB with exclusive option.
Activated volume group in Exclusive Mode.
Volume group "vg01PB" has been successfully changed.
Dec 1 07:29:13 - Node "psa1h01": Activating volume group vg01PD with exclusive option.
Activated volume group in Exclusive Mode.
Volume group "vg01PD" has been successfully changed.
Dec 1 07:29:13 - Node "psa1h01": Activating volume group vg01PL with exclusive option.
Activated volume group in Exclusive Mode.
Volume group "vg01PL" has been successfully changed.
Dec 1 07:29:13 - Node "psa1h01": Activating volume group vg01PS with exclusive option.
Activated volume group in Exclusive Mode.
Volume group "vg01PS" has been successfully changed.
Dec 1 07:29:14 - Node "psa1h01": Checking filesystems:
/dev/vg01PA/lvARCH1
/dev/vg01PB/lvBCKP1
/dev/vg01PD/lvDATA1
/dev/vg01PL/lvLOGS1
/dev/vg01PD/lvATCH1
/dev/vg01PD/lvPSHOME1
/dev/vg01PS/lvAUDIT1
/dev/vg01PA/rlvARCH1:file system is clean - log replay is not required
/dev/vg01PB/rlvBCKP1:file system is clean - log replay is not required
/dev/vg01PD/rlvDATA1:file system is clean - log replay is not required
/dev/vg01PL/rlvLOGS1:file system is clean - log replay is not required
/dev/vg01PD/rlvATCH1:file system is clean - log replay is not required
/dev/vg01PD/rlvPSHOME1:file system is clean - log replay is not required
/dev/vg01PS/rlvAUDIT1:file system is clean - log replay is not required
Dec 1 07:29:17 - Node "psa1h01": Mounting /dev/vg01PA/lvARCH1 at /oraarch01


Normally it startup right away... how can I find what happen?

Can I break this and return my package on the other node?

Jonathan

[Wouter Jagers]

Hiya,

Strange, since it does activate the VG and the fsck is ok.

Everything ok with the mountpoint ? "ls -ld /oraarch01" doesn't show anything weird on psa1h01 ?

Have you tried manually mounting it in a shell ?

The syslog might also mention something.

Cheers,
Wout

[skt_skt]

Check if yu had any change for VG on first node which might have not updated on another node(specially any disk addition.

Also check if you have any un-noticed errors on package cntl file..(specially the LV[x} numbering and mount options)

[Mridul Shrivastava]

whats the current status of the server now.. please share the information and let us know the status

[Jonathan Caplette_1]

Thanks you all..

I had to reboot the box...

Sorry for the late answer from me... My problem was because this VG is used in a SG environement... This is used with SAN disk and TrueCopy ... The problem is when I used to try running my package on second node, my SG script didn't do the takeover from TrueCopy... so this server tried to mount filesystems with no write permission on it...

I finally did the takeover manually and was able to mount the FS...

thanks again
Jonathan