Bug 1394773

Summary:	[Perf] : Getting EBADFD in brick logs while running I/O from multiple clients on Ganesha mounts.
Product:	Red Hat Gluster Storage	Reporter:	Ambarish <asoman>
Component:	nfs-ganesha	Assignee:	Kaleb KEITHLEY <kkeithle>
Status:	CLOSED DUPLICATE	QA Contact:	Sachin <sashinde>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	rhgs-3.2	CC:	asoman, bturner, kdhananj, kkeithle, mzywusko, pkarampu, rhinduja, rhs-bugs, skoduri, storage-qa-internal
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:	1394702	Environment:
Last Closed:	2016-11-16 07:12:35 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Bug Depends On:
Bug Blocks:	1394702

Comment 2 Krutika Dhananjay 2016-11-15 11:21:19 UTC

If you look at the following log from the bug report ...

bricks/bricks-testvol_brick3.log:[2016-11-13 18:12:07.130134] E [MSGID: 113072] [posix.c:3380:posix_writev] 0-testvol-posix: write failed: offset 0, [Bad file descriptor]
bricks/bricks-testvol_brick3.log:[2016-11-13 18:12:07.130216] E [MSGID: 115067] [server-rpc-fops.c:1348:server_writev_cbk] 0-testvol-server: 683223: WRITEV 10073 (3bc2aeb8-d197-4f9b-b6d2-c10588e17ba5) ==> (Bad file descriptor) [Bad file descriptor]

... posix translator is where the error originated. Posix attempted a pwrite on an fd and it failed with EBADF.

From the man page:

EBADF  fd is not a valid file descriptor or is not open for writing.

The above only means that posix xl logged a genuine failure.
This should not be mistaken to be a bug in posix translator itself.

I see that the parent bug https://bugzilla.redhat.com/show_bug.cgi?id=1394702 talks about a number of different issues seen on the same setup.

It's quite possible that this log is the result of the combination of those bugs.

Moving the bug back to nfs-ganesha as the bug is clearly not in posix translator, nor should posix be suppressing genuine failures like this which are very useful for debugging.

-Krutika