Red Hat Bugzilla – Bug 461043
lockd not using settings in sysconfig/nfs
Last modified: 2009-07-28 15:20:53 EDT
Description of problem:
The same problem as described in bug 434795 was also introduced with Red Hat Enterprise Linux 4.7.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. On an NFS server, add LOCKD_TCPPORT= and LOCKD_UDPPORT settings to /etc/sysconfig/nfs (e.g., LOCKD_TCPPORT=4001 and LOCKD_UDPPORT=4001)
2. Reboot; run "rpcinfo -p" to observe which ports are used for nlockmgr.
"rpcinfo -p" shows that the nlockmgr service is not using the ports defined in /etc/sysconfig/nfs. If the server's firewall rules depend on this, the nlockmgr service may be unreachable.
The nlockmgr service should be using the LOCKD_TCPPORT and LOCKD_UDPPORT settings.
The problem corresponds with the change to /etc/init.d/nfslock introduced between nfs-utils-1.0.6-84 and nfs-utils-126.96.36.199. Installing the older script (which probes the lockd module with the appropriate parameters) treats the symptom.
As per bug 434795, the cause of the problem is that the sysctl's in the new version of the nfslock script are ineffective if the lockd module has not been loaded, because the settings (fs.nfs.nlm_tcpport and fs.nfs.nlm_udpport) do not exist yet.
Fixed in nfs-utils-1.0.6-90.EL4
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.
*** Bug 496121 has been marked as a duplicate of this bug. ***