freebsd stable - Mar 2018 - Re: NFS 4.1 RECLAIM_COMPLETE FS failed error in combination with ESXi client

NAGY Andreas wrote:
>Thanks, the not issuing delegation warnings disappeared with this patch.
>
>But now there are some new warnings I haven't seen so far:
>2018-03-10T13:01:39.441Z cpu8:68046)WARNING: NFS41: NFS41FSOpGetObject:2148:
Failed to >get object 0x43910e71b386 [36 c6b10167 9b157f95 5aa100fb 8ffcf2c1
c 2 9f22ad6d 0 0 0 0 0]: >Stale file handle
I doubt these would be related to the patch. A stale FH means that the client
tried to
access a file via its FH after it was removed. (Normally this is a client bug,
but hopefully
not one that will cause grief.)
>These only appear several times after a the NFS share is mounted or
remounted after a >connection loss.
>Everything works fine, but haven't seen them till I applied the last
patch.
>
>andi
Ok. Thanks for testing all of these patches. I will probably get cleaned up
versions of
them committed in April.

The main outstanding issue is the Readdir one about directory changing too much.
Hopefully I can find out something about it via email.

Have fun with it, rick

Thanks! Please keep me updated if you find put more or when a updated version is
available.

As I now know it is working, I will start tomorrow to build up a testsystem with
3 NFS servers (two of them in a ha with CARP and HAST) and several ESXi hosts
which will all access there NFS datastores over 4 uplinks with NICs on different
subnets.
It should always be possible to do there some testing.

andi


________________________________
Von: Rick Macklem <rmacklem at uoguelph.ca>
Gesendet: 10.03.2018 11:20 nachm.
An: NAGY Andreas; 'freebsd-stable at freebsd.org'
Betreff: Re: NFS 4.1 RECLAIM_COMPLETE FS failed error in combination with ESXi
client

NAGY Andreas wrote:
>Thanks, the not issuing delegation warnings disappeared with this patch.
>
>But now there are some new warnings I haven't seen so far:
>2018-03-10T13:01:39.441Z cpu8:68046)WARNING: NFS41: NFS41FSOpGetObject:2148:
Failed to >get object 0x43910e71b386 [36 c6b10167 9b157f95 5aa100fb 8ffcf2c1
c 2 9f22ad6d 0 0 0 0 0]: >Stale file handle
I doubt these would be related to the patch. A stale FH means that the client
tried to
access a file via its FH after it was removed. (Normally this is a client bug,
but hopefully
not one that will cause grief.)
>These only appear several times after a the NFS share is mounted or
remounted after a >connection loss.
>Everything works fine, but haven't seen them till I applied the last
patch.
>
>andi
Ok. Thanks for testing all of these patches. I will probably get cleaned up
versions of
them committed in April.

The main outstanding issue is the Readdir one about directory changing too much.
Hopefully I can find out something about it via email.

Have fun with it, rick