help: return_layout status: -10025 - NFS4ERR_BAD_STATEID

help: return_layout status: -10025 - NFS4ERR_BAD_STATEID

[Boaz Harrosh]

Benny hi.

Only With Panfs. On a pnfs-2.6.33 Kernel after very heavy IO (git clone linux)
on the client side I get:

Jun  7 14:30:42 tl2 kernel: <-- return_layout status: -10025
Jun  7 14:30:42 tl2 kernel: <-- _pnfs_return_layout status: -10025

On the server side at the FS level I see a normal:
Jun  7 14:30:42 tl1 kernel: pan_kernel_fs_client_pnfs_layout_return: Begin I-xD02005194137f000f-xGe401b24c-xUb7850bdbb6117413 iomode=3 offs
et=0x0 length=0xffffffffffffffff cookie=0x0
Jun  7 14:30:42 tl1 kernel: pan_kernel_fs_client_pnfs_layout_return: released 3 caps
Jun  7 14:30:42 tl1 kernel: pan_kernel_fs_client_pnfs_layout_return: Return 0

These happen for 5 6 different files that do need return at that particular point.
(After Read was done and file closed by git)

And then everything at client side *freezes*.

1. Why would the Server return NFS4ERR_BAD_STATEID?
2. Why would the Client freeze and do nothing after that?

I'll try to run with nfs Server debug on to see what happens. Did Bruce fix
something like this for 2.6.34? or it was something else?

Boaz

Re: help: return_layout status: -10025 - NFS4ERR_BAD_STATEID

[Boaz Harrosh]

On 06/07/2010 02:51 PM, Boaz Harrosh wrote:
> Benny hi.
> 
> Only With Panfs. On a pnfs-2.6.33 Kernel after very heavy IO (git clone linux)
> on the client side I get:
> 
> Jun  7 14:30:42 tl2 kernel: <-- return_layout status: -10025
> Jun  7 14:30:42 tl2 kernel: <-- _pnfs_return_layout status: -10025
> 
> On the server side at the FS level I see a normal:
> Jun  7 14:30:42 tl1 kernel: pan_kernel_fs_client_pnfs_layout_return: Begin I-xD02005194137f000f-xGe401b24c-xUb7850bdbb6117413 iomode=3 offs
> et=0x0 length=0xffffffffffffffff cookie=0x0
> Jun  7 14:30:42 tl1 kernel: pan_kernel_fs_client_pnfs_layout_return: released 3 caps
> Jun  7 14:30:42 tl1 kernel: pan_kernel_fs_client_pnfs_layout_return: Return 0
> 
> These happen for 5 6 different files that do need return at that particular point.
> (After Read was done and file closed by git)
> 
> And then everything at client side *freezes*.
> 
> 1. Why would the Server return NFS4ERR_BAD_STATEID?
> 2. Why would the Client freeze and do nothing after that?
> 
> I'll try to run with nfs Server debug on to see what happens. Did Bruce fix
> something like this for 2.6.34? or it was something else?
> 
> Boaz
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

At the server it all starts when:
Jun  7 14:59:11 tl1 kernel: NFSD: laundromat service - starting
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12626)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12627)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12628)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12629)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12630)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12631)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12632)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12633)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12634)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12635)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12636)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12637)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12638)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12639)
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 12640)
<...>
Jun  7 14:59:11 tl1 kernel: NFSD: purging unused open stateowner (so_id 13080)
Jun  7 14:59:11 tl1 kernel: NFSD: laundromat_main - sleeping for 82 seconds

At that point the client gets these 10025, and freezes.

I'll try pnfs-2.6.34 just for fun.

Boaz