Hide Forgot
Description of problem: In our production system running Scientific 6 we experience problems with our nfs mount. Since the update to kernel 2.6.32-220 the dmesg log is filled with errors like NFS: directory brand-catcher/sessions contains a readdir loop. Please contact your server vendor. Offending cookie: 230704076 This usually happens around one time an hour. We don't notice a real failure - but one time the files of one directory was not accessible anymore and the dmesg was full of those messages. We have to webservers accessing a common nfs storage on a third server. Webserver1 is printing those errors since the update to the 220 kernel. Webserver2 is still using 2.6.32-131.17.1 and does not show the errors. That makes us quite sure that problem came with the update. So the problem is likely caused by the client. The nfs server itself is still running 2.6.32-71.29.1 . Additional info: There is a report of this problem for centos too: https://www.centos.org/modules/newbb/viewtopic.php?topic_id=35387 It's stated that this problem was introduced in kernel 2.6.38, fixed in 2.6.39 and may be backported to the recent redhat kernel.
Hi Bjorn, Please open a support ticket with Red Hat support. They can help diagnose and collect information and inform you of what issues have already been resolved. Thanks!
This was fixed in the RHEL6.2 update release....