On Mar 10, 2008 16:23 -0400, Peter Staubach wrote: > Here is version 3 of a patch set which modifies the system to > enhance the ESTALE error handling for system calls which take > pathnames as arguments. This patch set is essentially the > same as the v2 patches, but updated to reflect the current > state of the code around them. [snip long discussion of ESTALE causes] > This support was tested using the attached programs and > running multiple copies on mounted file systems which do not > share superblocks. When two or more copies of this program > are running, many ESTALE errors can be seen over the network. > Without these patches, the test program errors out almost > immediately. With these patches, the test program runs > for as long one desires. Have you tried "racer.sh"? That is a very stressful metadata tester that does random operations on a handful of file and directory names. It can be run on a single client, or on multiple clients and needs no coordination between the clients. I guess it won't tell you if you are getting ESTALE back correctly or not, but it can quickly find if there are any problems with the retrying code. I've attached an updated tarball of the original scripts here. Cheers, Andreas -- Andreas Dilger Sr. Staff Engineer, Lustre Group Sun Microsystems of Canada, Inc.
Attachment:
racer-lustre.tar.gz
Description: GNU Zip compressed data
