I'm seeing repeated duplicated queries going on between a NFSv4 client
and server. Both client and server are running linux v3.9.2, but I have
observed this with other versions on both.
Wireshark decodes it as (copied by hand):
V4 Call (Reply In 3) WRITE StateID:0x1c7c Offset:0 Len:8192
V4 Reply (Call In 2) WRITE Status: NFS4ERR_OLD_STATEID
I've put a tcpdump recording up here:
http://untroubled.org/nfsv4.tcpdump.xz
This behavior seems to be most reliably triggered by using Chromium, but
I have observed it continuing even after shutting that down. I am able
to continue to use the client while this behavior is going on, albeit
with considerably reduced NFS performance.
What is going on here?
--
Bruce Guenter <[email protected]> http://untroubled.org/
On Mon, May 13, 2013 at 02:22:20PM -0600, Bruce Guenter wrote:
> I'm seeing repeated duplicated queries going on between a NFSv4 client
> and server. Both client and server are running linux v3.9.2, but I have
> observed this with other versions on both.
>
> Wireshark decodes it as (copied by hand):
>
> V4 Call (Reply In 3) WRITE StateID:0x1c7c Offset:0 Len:8192
> V4 Reply (Call In 2) WRITE Status: NFS4ERR_OLD_STATEID
I am still seeing this behavior with the client running v3.10-rc6 and
server on v3.9.3. The last two times I've seen it, gvfsd-metadata has
been blocked in D state.
Does anybody have any idea what's going on here?
--
Bruce Guenter <[email protected]> http://untroubled.org/