Hi there,

OES 2 cluster: I updated multipathing (Dell PowerPath -> 5.1.1, x86_64) on both nodes by rebooting a node into single user mode, upgrading the PowerPath rpm package to the version above, and then rebooting into normal mode. Everything comes up OK except for namcd (and subsequent services: apache, tomcat, xsrvd etc) - the following is in /var/log/messages:

Aug 12 14:56:58 node1a kernel: namcd[6560]: segfault at 00000000005c4000 rip 00000000004170a4 rsp 00007fffb242ebc8 error 4

Unusual. But if I then type "namconfig cache_refresh" and then start the subsequent services manually they all come up fine (including NSS drives and NCS clustering, and I can failover volumes) - but the segfault error still persists on reboot and namconfig needs to be manually run with cache_refresh.

To try and fix I did "namconfig rm" and reinstalled namcd (LUM) through the OES2 Yast install/configure. This fixes the segfault issue, but it still won't start services on reboot, it can't seem to pull down the right users from eDirectory (wwwrun, etc). A "getent passwd" command only shows me an "admin" user from eDirectory, but no wwwrun or novlxsrvd user. If I do "namuserlist -x o=maintree" however, I can see wwwrun users.

If I manually restart namcd and look at /var/log/messages, I get the following:

Aug 14 15:43:37 node1a /usr/sbin/namcd[16735]: deinitialized the worker threads
Aug 14 15:43:37 node1a /usr/sbin/namcd[16735]: deinitialized the cache refresh thread
Aug 14 15:43:37 node1a /usr/sbin/namcd[16735]: monitorChangesInLDAP: ldap_result: Can't contact LDAP server
Aug 14 15:43:40 node1a /usr/sbin/namcd[16735]: deinitialized the LDAP watcher thread
Aug 14 15:43:42 node1a /usr/sbin/namcd[16735]: Deleted hash tables and flushed data into local files
Aug 14 15:43:42 node1a /usr/sbin/namcd[16735]: Deinitialized threads
Aug 14 15:43:44 node1a namcd: SIGTTOU caught
Aug 14 15:43:44 node1a namcd: SIGTTIN caught
Aug 14 15:43:44 node1a namcd: SIGTSTP caught
Aug 14 15:43:44 node1a /usr/sbin/namcd[16900]: Starting namcd..
Aug 14 15:43:44 node1a /usr/sbin/namcd[16900]: namcd populating the user hash tables
Aug 14 15:43:44 node1a /usr/sbin/namcd[16900]: namcd populating group hash tables
Aug 14 15:43:44 node1a /usr/sbin/namcd[16900]: namcd Populated hash tables
Aug 14 15:43:44 node1a /usr/sbin/namcd[16900]: Created all the threads

Anyone any ideas? Does it look like there is something wrong with LDAP with the message "Can't contact LDAP server"? Obviously the server is contactable though, all other services (edir etc) are working fine. And it appears namcd does pull down the "admin" user from edirectory, just not the wwwrun or novlxsrvd users etc.

Do I need to reconfigure services?

Thanks for any ideas you have,

Karl