Sun Fire[TM] 12K/15K/20K/25K: cfgadm -c disconnect fails with Library error: RCM request delete capacity failed for SB

Asset ID:	1-72-1002437.1
Update Date:	2015-12-02
Keywords:

Solution Type Problem Resolution Sure

Solution 1002437.1 : Sun Fire[TM] 12K/15K/20K/25K: cfgadm -c disconnect fails with Library error: RCM request delete capacity failed for SB

Applies to:

Sun Fire 12K Server - Version Not Applicable and later
Sun Fire 15K Server - Version Not Applicable and later
Sun Fire E20K Server - Version Not Applicable and later
Sun Fire E25K Server - Version Not Applicable and later
All Platforms

Symptoms

When running cfgadm -c disconnect SB#, following error may occur:

Library error: RCM request delete capacity failed for SB#

Place the dr session into debug

SBD_DEBUG=/tmp/sbd.log;export SBD_DEBUG

and rerun the cfgadm -c disconnect SB#; looking at the /tmp/sbd.log shows we are failing with a bad address:

....snip ....
<request delete capacity><SB?>ap_rcm_info(PID)
<Bad address><><>
ap_seq_exec: request delete capacity failed
.... snip ....

Cause

This error will occur if /var/run is not mounted as tmpfs.

Solution

Using mount -p /var/run should be mounted as a tmpfs filesystem.

# mount -p
......
swap - /var/run tmpfs - no xattr
......

Running Solaris[TM] 8 and Solaris[TM] 9, /var/run is mounted in the /etc/rcS.d/S70buildmnttab.sh, while /var/run is mounted via svc:/system/filesystem/minimal:default in Solaris[TM] 10.

Note: This problem can only happen with Solaris 8 and Solaris 9 because in Solaris 10 the mounting of /var/run is done much earlier and the check for /var/run in librcm has been removed (see Bug 15227924).

Internal Section

Keywords: rcm, cfgadm, disconnect, deleteboard

Previously Published As 83347

References

<BUG:15223544> - SUNBT5088931 DELETEBOARD CMD RESULTS IN RCM LIBRARY ERROR
<BUG:15227924> - SUNBT5103374-SOLARIS_11 /VAR/RUN CHECK IN LIBRCM SHOULD BE REMOVED

Attachments

This solution has no attachment