LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* RPC request reserved 0 but used 96
@ 2004-05-15  8:38 Jens Axboe
  2004-05-15  8:58 ` Jens Axboe
  0 siblings, 1 reply; 4+ messages in thread
From: Jens Axboe @ 2004-05-15  8:38 UTC (permalink / raw)
  To: Linux Kernel

Hi,

Seeing lots of these on a small server that hosts nfs shares (root and
"normal").

router:~ # dmesg | tail -n5
RPC request reserved 0 but used 96
RPC request reserved 0 but used 96
RPC request reserved 0 but used 140
RPC request reserved 0 but used 140
RPC request reserved 0 but used 96

I see nfs stalls on the client, doesn't seem to be directly related to
when the above messages happen.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: RPC request reserved 0 but used 96
  2004-05-15  8:38 RPC request reserved 0 but used 96 Jens Axboe
@ 2004-05-15  8:58 ` Jens Axboe
  2004-05-15 10:36   ` Jens Axboe
  0 siblings, 1 reply; 4+ messages in thread
From: Jens Axboe @ 2004-05-15  8:58 UTC (permalink / raw)
  To: Linux Kernel; +Cc: Trond Myklebust, Andrew Morton

On Sat, May 15 2004, Jens Axboe wrote:
> Hi,
> 
> Seeing lots of these on a small server that hosts nfs shares (root and
> "normal").
> 
> router:~ # dmesg | tail -n5
> RPC request reserved 0 but used 96
> RPC request reserved 0 but used 96
> RPC request reserved 0 but used 140
> RPC request reserved 0 but used 140
> RPC request reserved 0 but used 96
> 
> I see nfs stalls on the client, doesn't seem to be directly related to
> when the above messages happen.

This went out a little early, lots of pieces missing.

The server is running 2.6.6-mm2, NFS options are as follows:

CONFIG_NFS_FS=y
CONFIG_NFS_V3=y
# CONFIG_NFS_V4 is not set
CONFIG_NFS_DIRECTIO=y
CONFIG_NFSD=y
CONFIG_NFSD_V3=y
# CONFIG_NFSD_V4 is not set
CONFIG_NFSD_TCP=y
CONFIG_LOCKD=y
CONFIG_LOCKD_V4=y
CONFIG_EXPORTFS=y
CONFIG_SUNRPC=y

The shares are exported async.

Client is running 2.4.26, mount options are nfsvers=3,tcp.

I'll be trying 2.6.6-BK on the server now.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: RPC request reserved 0 but used 96
  2004-05-15  8:58 ` Jens Axboe
@ 2004-05-15 10:36   ` Jens Axboe
       [not found]     ` <20040518040613.GB10633@fieldses.org>
  0 siblings, 1 reply; 4+ messages in thread
From: Jens Axboe @ 2004-05-15 10:36 UTC (permalink / raw)
  To: Linux Kernel; +Cc: Trond Myklebust, Andrew Morton

On Sat, May 15 2004, Jens Axboe wrote:
> On Sat, May 15 2004, Jens Axboe wrote:
> > Hi,
> > 
> > Seeing lots of these on a small server that hosts nfs shares (root and
> > "normal").
> > 
> > router:~ # dmesg | tail -n5
> > RPC request reserved 0 but used 96
> > RPC request reserved 0 but used 96
> > RPC request reserved 0 but used 140
> > RPC request reserved 0 but used 140
> > RPC request reserved 0 but used 96
> > 
> > I see nfs stalls on the client, doesn't seem to be directly related to
> > when the above messages happen.
> 
> This went out a little early, lots of pieces missing.
> 
> The server is running 2.6.6-mm2, NFS options are as follows:
> 
> CONFIG_NFS_FS=y
> CONFIG_NFS_V3=y
> # CONFIG_NFS_V4 is not set
> CONFIG_NFS_DIRECTIO=y
> CONFIG_NFSD=y
> CONFIG_NFSD_V3=y
> # CONFIG_NFSD_V4 is not set
> CONFIG_NFSD_TCP=y
> CONFIG_LOCKD=y
> CONFIG_LOCKD_V4=y
> CONFIG_EXPORTFS=y
> CONFIG_SUNRPC=y
> 
> The shares are exported async.
> 
> Client is running 2.4.26, mount options are nfsvers=3,tcp.
> 
> I'll be trying 2.6.6-BK on the server now.

2.6.6-BK with shares exported sync show the same behaviour. I get:

RPC request reserved 0 but used 32900
RPC request reserved 0 but used 32900
RPC request reserved 0 but used 96

and client stalls with a:

nfs: server router not responding, still trying
nfs: server router OK

at that time.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: RPC request reserved 0 but used 96
       [not found]     ` <20040518040613.GB10633@fieldses.org>
@ 2004-05-18 11:03       ` Jens Axboe
  0 siblings, 0 replies; 4+ messages in thread
From: Jens Axboe @ 2004-05-18 11:03 UTC (permalink / raw)
  To: J. Bruce Fields; +Cc: Linux Kernel, Trond Myklebust, Andrew Morton

On Tue, May 18 2004, J. Bruce Fields wrote:
> On Sat, May 15, 2004 at 12:36:51PM +0200, Jens Axboe wrote:
> > 2.6.6-BK with shares exported sync show the same behaviour. I get:
> > 
> > RPC request reserved 0 but used 32900
> > RPC request reserved 0 but used 32900
> > RPC request reserved 0 but used 96
> 
> Does this end the messages?
> 
> --Bruce Fields
> 
> svc_recv may call svc_sock_release before rqstp->rq_res is initialized.
> 
>  net/sunrpc/svcsock.c |    1 +
>  1 files changed, 1 insertion(+)
> 
> diff -puN net/sunrpc/svcsock.c~svc_reserve_debugging net/sunrpc/svcsock.c
> --- linux-2.6.6/net/sunrpc/svcsock.c~svc_reserve_debugging	2004-05-17 23:46:51.000000000 -0400
> +++ linux-2.6.6-bfields/net/sunrpc/svcsock.c	2004-05-17 23:50:54.000000000 -0400
> @@ -1255,6 +1255,7 @@ svc_recv(struct svc_serv *serv, struct s
>  
>  	/* No data, incomplete (TCP) read, or accept() */
>  	if (len == 0 || len == -EAGAIN) {
> +		rqstp->rq_res.len = 0;
>  		svc_sock_release(rqstp);
>  		return -EAGAIN;
>  	}

Seems to have fixed it, at least I haven't seen any messages of this
sort since applying the patch and switching back to nfs over tcp.
Thanks!

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2004-05-18 11:03 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-05-15  8:38 RPC request reserved 0 but used 96 Jens Axboe
2004-05-15  8:58 ` Jens Axboe
2004-05-15 10:36   ` Jens Axboe
     [not found]     ` <20040518040613.GB10633@fieldses.org>
2004-05-18 11:03       ` Jens Axboe

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).