Red Hat Bugzilla – Bug 823880
[glusterfs-3.3.0-qa42]: null gfid id is printed if symlink fails
Last modified: 2013-07-24 14:04:03 EDT
Description of problem:
In client-protocol suppose symlink fails, then in the log message null gfid is printed. We should not try to print the gfid since inode will not be containing the gfid because of the failure.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. create a volume and mount it
2. run a test where symlink is failed.
null gfid is printed in the log
null gfid should not be printed since symlink is failed and inode will not be containing any gfid.
[2012-05-22 15:05:21.816002] W [client3_1-fops.c:187:client3_1_symlink_cbk] 1-mirror-client-2: remote operation failed: File name too
long. Path: /run25872/p7/l2 (00000000-0000-0000-0000-000000000000)
http://review.gluster.com/3411 fixes the issue and null gfid is not printed if symlink operation fails.
I just tested on the 3.4qa8 and when I run ltp I am still seeing:
I also see this error on the latest 3.3 qa bits. I am assuming that you are running ltp based on the output from your error message, is that correct? Are you also seeing errors like this in the logs or am I seeing a different issues than htis BZ was opened for?
I think this log is ok. This is logged by the posix xlator as the name of the symlink given for creation was larger than the posix standard. Here you can see that NULL gfid is not logged. In the earlier case, when a this kind of things happen client xlator was logging the NULL gfid apart from the name of the entry. So the logging of NULL gfid is removed now in the client xlator.