systemimager-commits

[<ar...@sy...>]

Author: arighi
Date: 2008-04-12 08:19:40 -0500 (Sat, 12 Apr 2008)
New Revision: 4442

Modified:
   trunk/etc/getimage.exclude
Log:
dh...@jl... wrote:

 I started seeing these error messages on all of our Fedora 8 imaged machines:

 Mar 31 11:22:02 <HOST> rpc.statd[2017]: creat(/var/lib/nfs/statd/sm/<HOST>) failed: No such file or directory
 Mar 31 11:22:02 <HOST> rpc.statd[2017]: STAT_FAIL to <HOST> for SM_MON of <IP ADDR>
 Mar 31 11:22:02 <HOST> kernel: lockd: cannot monitor <HOST>

 Turns out that this is present in getimage.exclude by default in 4.0.2:

 # NFS stuff
 /var/lib/nfs/state
 /var/lib/nfs/*tab
 /var/lib/nfs/*/*

 I'm fine with NFS data being removed during the imaging process, but destroyed
 static dirs should at least be replaced. /var/lib/nfs/statd/sm is owned by
 rpcuser and is part of the nfs-utils rpm package. Since it is not automatically
 recreated rpc spews tons of error messages when it can't write it's metadata.
 mkdir fixes the issue and is simple to fix, but if possible it would be nice to
 not have the sm dir destroyed by the default SI install.

This check-in fixes the bug above explicitly including /var/lib/nfs/statd/sm
(and /var/lib/nfs/statd/sm.bak) in the image.

Bug reported by: <dh...@jl...>


Modified: trunk/etc/getimage.exclude
===================================================================
--- trunk/etc/getimage.exclude	2008-04-12 10:53:19 UTC (rev 4441)
+++ trunk/etc/getimage.exclude	2008-04-12 13:19:40 UTC (rev 4442)
@@ -8,9 +8,12 @@
 /media/*
 
 # NFS stuff
++ /var/lib/nfs/statd/sm
++ /var/lib/nfs/statd/sm.bak
 /var/lib/nfs/state
 /var/lib/nfs/*tab
 /var/lib/nfs/*/*
+/var/lib/nfs/*/*/*
 
 # LVM caches and backups (automatically re-created at the first boot)
 /etc/lvm/.cache