HELP with permissions *** SAM & Service Guard

[Alessio_6]

Hi,

there has been some mess about /usr permissions, so strange thins are happening:
1) SAM doesn't work: it returns:

UNEXPECTED EXIT: process LANG=C LC_ALL=C /usr/sam/lbin/samx -C -p UNEXPECTED EXIT: process LANG=C LC_ALL=C /usr/sam/lbin/samx -C -p status
you must have superuser (root) privilege to enter SAM


2) our service guard cluster doesn't form anymore on this node, returning the messages:

Sep 26 18:23:50 dapct2 CM-CMD[3311]: cmruncl -v
Sep 26 18:23:50 dapct2 cmclconfd[3313]: Executing "/usr/lbin/cmcld" for node dapct2
Sep 26 18:23:50 dapct2 cmcld: Daemon Initialization - Maximum number of packages supported for this incarnation is 5.
Sep 26 18:23:50 dapct2 cmcld: Global Cluster Information:
Sep 26 18:23:50 dapct2 cmcld: Heartbeat Interval is 1 seconds.
Sep 26 18:23:50 dapct2 cmcld: Node Timeout is 8 seconds.
Sep 26 18:23:50 dapct2 cmcld: Network Polling Interval is 2 seconds.
Sep 26 18:23:50 dapct2 cmcld: Auto Start Timeout is 600 seconds.
Sep 26 18:23:51 dapct2 cmcld: Information Specific to node dapct2:
Sep 26 18:23:51 dapct2 cmcld: Cluster lock disk: /dev/dsk/c4t9d0.
Sep 26 18:23:51 dapct2 cmcld: lan0 0x00306e1bcf51 172.16.14.10 bridged net:1
Sep 26 18:23:51 dapct2 cmcld: lan1 0x00306e0c961f standby bridged net:1
Sep 26 18:23:51 dapct2 cmcld: lan2 0x00306e0c8689 192.0.100.2 bridged net:2
Sep 26 18:23:51 dapct2 cmcld: Heartbeat Subnet: 172.16.14.0
Sep 26 18:23:51 dapct2 cmcld: The maximum # of concurrent local connections to the daemon that will be supported is 33.
Sep 26 18:23:51 dapct2 cmlogd: Could not set realtime priority: Not owner
Sep 26 18:23:51 dapct2 cmlvmd: Unable to lock into memory (Not owner)
Sep 26 18:23:51 dapct2 cmlvmd: Startup failed.
Sep 26 18:24:01 dapct2 cmcld: Processing exit status for service cmlogd
Sep 26 18:24:01 dapct2 cmcld: Service cmlogd terminated due to an exit(1).
Sep 26 18:24:01 dapct2 cmcld: Automatically restarted service cmlogd for the 1st time after failure.
Sep 26 18:24:01 dapct2 cmcld: Processing exit status for service cmlvmd
Sep 26 18:24:01 dapct2 cmcld: Service cmlvmd terminated due to an exit(1).
Sep 26 18:24:01 dapct2 cmcld: Daemon exiting to preserve data integrity
Sep 26 18:24:01 dapct2 cmcld: Reason: LVM daemon failed
Sep 26 18:24:01 dapct2 cmsrvassistd[3317]: The cluster daemon aborted our connection.
Sep 26 18:24:01 dapct2 cmsrvassistd[3317]: Lost connection with ServiceGuard cluster daemon (cmcld): Software caused connection abort
Sep 26 18:24:01 dapct2 cmsrvassistd[3323]: The cluster daemon aborted our connection.
Sep 26 18:24:01 dapct2 cmsrvassistd[3323]: Unable to notify ServiceGuard main daemon (cmcld): Software caused connection abort


I tried to set the right permissions, by comparing them with a good /usr dir, but I get the same thing.
Could you help me?


thanks...
Alessio

[RAC_1]

I assume you ran those command with root id
to confirm check id, it should show uid 0.

What permissions you have for sam.
it shoud be bin:bin for usr
For cmruncl it shoud be the same.

[Jeff Schussele]

Hi Alessio,

Couple of things.
Here's the proper perms for /usr/lbin/cm*

# ll cm*
-r-xr--r-- 1 bin bin 1378000 Aug 8 2001 cmclconfd
-r-xr--r-- 1 bin bin 3344080 Aug 8 2001 cmcld
-r-xr--r-- 1 bin bin 163840 Aug 8 2001 cmlogd
-r-xr--r-- 1 bin bin 960208 Aug 8 2001 cmlvmd
-r-xr--r-- 1 bin bin 2076672 Aug 8 2001 cmsnmpd
-r-xr--r-- 1 bin bin 177872 Aug 8 2001 cmsrvassistd
-r-xr--r-- 1 bin bin 1042128 Aug 8 2001 cmtaped
-r-xr-xr-x 1 bin bin 3176144 Aug 8 2001 cmui

And for /usr/lbin itself
dr-xr-xr-x 11 bin bin 2048 Aug 27 22:13 lbin

Here's perms for /usr/sam/lbin/sam*

# ll sa*
-r-xr-xr-x 1 bin bin 16384 Jun 21 1999 sam_nis_ismember
-r-xr-xr-x 1 bin bin 16384 Jun 21 1999 sam_niscat
-r-xr-xr-x 1 bin bin 16384 Jun 21 1999 sam_nisgrp
-r--r--r-- 1 bin bin 15294 Oct 27 1997 sam_utils
-r-xr-xr-x 1 bin bin 307200 Nov 28 2000 samagent
-r-xr-xr-x 1 bin bin 24576 Oct 27 1997 samlog
-r-xr-xr-x 1 bin bin 290816 Jun 8 2001 samx

But other things I would check are - what's root's UID in /etc/passwd? Is it still 0?
If trusted, is root's entry in the TCB good?
Check your /etc/passwd file for corruption - run pwck to verify the passwd file integrity.

If you run trusted - I'd also check the TCB with
/usr/sbin/authck -av
for any problems and/or corruption.

HTH,
Jeff

[melvyn burnard]

run swverify and use it's output log file to check and correct permissions/ownerships.
If that does not help, recover to a known good backup.
If all else fails, you may have to re-install, but the first step is the swverify