Error 41 - bpbkar using 98% cpu

This Wiki is brought to you by Backup Central, where you can find the Mr. Backup Blog, Forums, and a mailing list for each forum!
Backup FAQs	Service Providers	Backup Software	Backup Hardware	Backup Book Wiki	Free Stuff	Miscellaneous

The manual backup job works but the scheduled ones die with an error 41 and leave 3 bpbkar processes hanging and sucking up cpu resources.

You go in and manually kill the bpbkar processes that are hung but the next day it does it again and fails all 3 attempts leaving 3 more processes in limbo.

This may be related to vold not behaving properly. This can occur if you use vold to mount CDROMs and then kill it afterwards. Even though the process isn't running it still has a lock on a directory (usually /vol or /cdrom). The process vold appears to have an NFS server inside it (see Sun patch 102081).

A second symptom is that trying to look at the /vol filesystem hangs up:

1. cd /
2. ls -al

NFS server for volume management (/vol) not responding still trying ^CNFS getattr failed for server for volume management (/vol): error 23 (RPC: Unitdata error)

1. umount /vol

nfs umount: cbaicidcweb01:vold(pid2457) server not responding: RPC: Rpcbind failure - RPC: Unable to receive

1. umount /cdrom

umount: warning: /cdrom not in mnttab umount: /cdrom not mounted

Once you can do a "ls -al" in the root directory and can see /vol it should work.

---

SUMMARY: vold dies and won't restart PROBLEM DESCRIPTION

From time to time vold crashes and it will not restart.

Mon Jan 29 17:55:30 2001 fatal: mounting of "/vol" failed A second symptom is that trying to look at the /vol filesystem hangs up:

% ls /vol NFS server for volume management (/vol) not responding still trying it can take tens of minutes to time out.

SOLUTION This worked for me, but it will probably not cure all the ills of vold. I assume you do not have vold running, but you _want_ it to be running.

1. check /etc/mnttab for old mounts belonging to vold: % grep vol /etc/mnttab attila:vold(pid2058) /vol nfs ignore,noquota,dev=2b807f6981095587 % grep cdrom /etc/mnttab

Vold appears to have an NFS server inside it (see Sun patch 102081) and it seems that it will not restart if these entries are present in /etc/mnttab.

The error messages are not very helpful in this regard - they don't give any clues _why_ /vol cannot be mounted.

2. Clean out the mnttab by umounting the relevant mount points

1. umount /cdrom/cdrom
2. umount /vol

(you may need to unshare these first if they are nfs-exported; check in /etc/dfs/sharetab)

2a. I did this, but don't know if it matters

1. rm /tmp/.removable/cdrom1

This directory appears to be used by vold to keep track of mount points.

3. Restart vold

1. /etc/init.d/volmgt start

BTW: If you are debugging vold problems, try running in debug mode;

1. /usr/sbin/vold -t -v -L 99 -l /var/adm/vold.debug_log

it didn't help in this case but it may for others.