Description of problem: bz 170714 introduced a fix to allow more nfs mounts to be established. This solution however, assumed the availability of util-linux-2.11y-31.14, which contains a fix to allow mount to always use UDP first when contact the mountd rpc program daemon. Without this fix in place, the fix in bz 170714 actually causes a regression in the number of supported mounts over nfs. Version-Release number of selected component (if applicable): How reproducible: always Steps to Reproduce: 1.attempt to mount 1024 nfs mounts under RHEL3 U4 and RHEL3 U7 2. 3. Actual results: RHEL3 U4 mounts >= 800 nfs mount RHEL3 U7 mounts == ~512 nfs mounts Expected results: Additional info:
Removing CANFIX blocker... this has not gone through ACK process.
Hmm.. there is some unreleased patch: * Wed Oct 19 2005 Steve Dickson <steved> 2.11y-31.13 - fix #170714 Changed nfsmount to first try the UDP transport when querying the remote rpc.mountd, then tcp iff tcp is specified. This tremendously increases the total number of tcp mounts that can occur. I think (hope) this patch resolve the problem. It means we needn't rebase anything.
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 the 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. http://rhn.redhat.com/errata/RHBA-2006-0456.html