Bad NFS performance if share is loopback mounted
This document (7009907) is provided subject to the disclaimer at the end of this document.
Environment
All Product versions
All Servicepacks
Situation
This looks like to work for normal conditions, but sometimes it is possible to experience bad performance during read or write to the shares. This bad performance is seen on the NFS server itself and/or on other NFS clients. The bad performance can go up to and including timeouts, locks and even deadlocks.
Generally we do NOT recommend or support this "loopback" NFS mounts.
Resolution
To use a bindmount simply use the mount option "bind"
mount -o bind /myexport/myshare /clientmountpointinstead of
mount -t nfs <myip>:/myexport/myshare /clientmountpoint
or
mount -t nfs 127.0.0.1:/myexport/myshare /clientmountpoint
The entry for a bindmount in /etc/fstab would look like
/myexport/myshare /clientmountpoint none bind 0 0
Additional Information
- NFS server needs more memory
- NFS server tries to flush page cache
- page cache tries to free memory by flushing page cache allocated by files over NFS
- NFS client tries to flush pages by writing back to NFS server. Go back to 1.
Disclaimer
This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.
- Document ID:7009907
- Creation Date: 19-Dec-2011
- Modified Date:03-Mar-2020
-
- SUSE Linux Enterprise Desktop
- SUSE Linux Enterprise Point of Service
- SUSE Linux Enterprise Real Time
- SUSE Linux Enterprise Server
- SUSE Studio Onsite
- SUSE Lifecycle Management Server
For questions or concerns with the SUSE Knowledgebase please contact: tidfeedback@suse.com