I wanted to see if you guys had any ideas about a problem I'm having.
Usually things work just fine, but occasionally
This problem only crops up when the file server is under load and even
then randomly. I suspect the problem is in the ARC metadata, but I'm
not sure. When the CIFS client makes a call:
FASTIO_QUERY_OPEN the result is: FAILURE
IRP_MJ_CREATE the result is: NO SUCH FILE
Of course these files do exist and are accessible to any client. The
solution has been to open the folder in a windows CIFS client, deselect
the Read-Only attribute and apply that. This unsets the Read-Only
attribute which (I've verified) was not set on any object in the
directory in the first place. Once that happens everything works just
fine until some point in the future when the process happens again.
One other thing to note, the file server has over 30K users and stores
several million files, but it is generally the same paths that this
issue continues to happen to. The users access the share via a common
process (IMAP server). Also this behavior does not happen on our older
7110 storage appliance.
Has this behavior been seen before, and are there any suggestions to
fix/debug further?
Thanks all!
_______________________________________________
cifs-discuss mailing list
cifs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/cifs-discuss