NFS, Can't get request slot

NFS, Can't get request slot

[Grahame Jordan]

[-- Attachment #1: Type: text/plain, Size: 1098 bytes --]

Hi,

We have /usr mounted over NFS on our workstations  RH6.2
Server RH 6.2
nfs-utils-0.1.9.1-1
Kernel 2.2.16

These workstations happily use samba and other services without any
delays but with NFS they hang in X for up to 15 minutes before NFS come
back.
We can ssh into the workstations and use any utility underneath X
without any problems whilst it is hung in X.

We have changed  the Server eepro100 for a 3com 3c95x with no difference

according to what has been alluded to in other kernel posts.

By the evidence that we have gathered it seems that the Server is not
taxed too much as samba users are getting files OK etc.  The can't get
request slot is plaguing many others in different ways.   It looks like
an NFS issue.   How can this be proven?  Then we can work on the
problem.

Thanks

--
Grahame Jordan
Network Manager
Interim Technology Training Institute
  Mobile: +61 3 0408 058 209
  Phone:  +61 3 9243 2220
  Fax:    +61 3 9820 2010
  e-mail: jordg@cpgen.cpg.com.au
  Transforming the way people work with technology with
  INTEGRITY LEARNING INNOVATION TEAMWORK PERFORMANCE



[-- Attachment #2: Type: text/html, Size: 1355 bytes --]

Re: NFS, Can't get request slot

[Alan Cox]

> By the evidence that we have gathered it seems that the Server is not
> taxed too much as samba users are getting files OK etc.  The can't get
> request slot is plaguing many others in different ways.   It looks like
> an NFS issue.   How can this be proven?  Then we can work on the
> problem.

The request queue slot message means the server isnt responding (at least in
the eyes of the client). Given you can get into the box that isnt the
net card (at least not now). What mount options do you use ?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: NFS, Can't get request slot

[Grahame Jordan]

[-- Attachment #1: Type: text/plain, Size: 1139 bytes --]

Alan,

student:/susr/     /usr              nfs
defaults,rsize=1024,wsize=1024          0 0

Thanks

Graham Jordan


Alan Cox wrote:

> > By the evidence that we have gathered it seems that the Server is not
> > taxed too much as samba users are getting files OK etc.  The can't get
> > request slot is plaguing many others in different ways.   It looks like
> > an NFS issue.   How can this be proven?  Then we can work on the
> > problem.
>
> The request queue slot message means the server isnt responding (at least in
> the eyes of the client). Given you can get into the box that isnt the
> net card (at least not now). What mount options do you use ?
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> Please read the FAQ at http://www.tux.org/lkml/

--
Grahame Jordan
Network Manager
Interim Technology Training Institute
  Mobile: +61 3 0408 058 209
  Phone:  +61 3 9243 2220
  Fax:    +61 3 9820 2010
  e-mail: jordg@cpgen.cpg.com.au
  Transforming the way people work with technology with
  INTEGRITY LEARNING INNOVATION TEAMWORK PERFORMANCE



[-- Attachment #2: Type: text/html, Size: 1622 bytes --]

Re: NFS, Can't get request slot

[Grahame Jordan]

[-- Attachment #1: Type: text/plain, Size: 3962 bytes --]

Alan,

On a client running all by itself on our production network overnight, doing an
"ls -lR /usr" and doing a compile (in a loop) is giving the same problems.

A little snipit from /var/log/messages
Nov 10 04:31:20 spc81 kernel: nfs: server student OK
Nov 10 04:33:17 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:33:17 spc81 kernel: nfs: server student OK
Nov 10 04:34:59 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:35:04 spc81 kernel: nfs: task 20823 can't get a request slot
Nov 10 04:35:35 spc81 kernel: nfs: server student OK
Nov 10 04:35:35 spc81 kernel: nfs: server student OK
Nov 10 04:44:43 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:44:44 spc81 kernel: nfs: server student OK
Nov 10 04:45:56 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:46:00 spc81 kernel: nfs: task 20457 can't get a request slot
Nov 10 04:46:00 spc81 kernel: nfs: server student OK
Nov 10 04:46:03 spc81 kernel: nfs: server student OK
Nov 10 04:49:15 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:49:15 spc81 kernel: nfs: server student OK
Nov 10 04:49:25 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:52:21 spc81 kernel: nfs: server student OK
Nov 10 04:53:43 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:53:43 spc81 kernel: nfs: server student OK
Nov 10 04:54:01 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:54:02 spc81 kernel: nfs: task 64822 can't get a request slot
Nov 10 04:54:13 spc81 kernel: nfs: server student OK
Nov 10 04:54:15 spc81 kernel: nfs: server student OK
Nov 10 04:55:57 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:55:59 spc81 kernel: nfs: task 29484 can't get a request slot
Nov 10 04:56:21 spc81 kernel: nfs: server student OK
Nov 10 04:56:23 spc81 kernel: nfs: server student OK
Nov 10 04:58:15 spc81 kernel: nfs: server student not responding, still trying
Nov 10 04:58:21 spc81 kernel: nfs: task 1477 can't get a request slot
Nov 10 04:58:52 spc81 kernel: nfs: task 1478 can't get a request slot
Nov 10 04:59:15 spc81 kernel: nfs: server student OK

Shows that under no (or little) load this thing is failing.

We have 20 networks with ~100 machines on each with this same problem.  Only a
few at a time are using Linux.  We are trying to promote Linux as a great
operating system, but this is not helping.  Again the samba on win95 is puring
along nicely plus every other service but for NFS.

We have changed the NIC on this server to 3Com 3c90x for no change in status.
Clients are using  CNETPro120 200 and Intel eepro100

The servers are generally PIII500  512MB RAM.

We have tried mount flags rsize=1024, 8192, 128, with no change.

It looks like a server issue related to RPC.   But I am only guessing here.

Any more help is much appreciated.

Thanks

Grahame Jordan


Alan Cox wrote:

> > By the evidence that we have gathered it seems that the Server is not
> > taxed too much as samba users are getting files OK etc.  The can't get
> > request slot is plaguing many others in different ways.   It looks like
> > an NFS issue.   How can this be proven?  Then we can work on the
> > problem.
>
> The request queue slot message means the server isnt responding (at least in
> the eyes of the client). Given you can get into the box that isnt the
> net card (at least not now). What mount options do you use ?
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> Please read the FAQ at http://www.tux.org/lkml/

--
Grahame Jordan
Network Manager
Interim Technology Training Institute
  Mobile: +61 3 0408 058 209
  Phone:  +61 3 9243 2220
  Fax:    +61 3 9820 2010
  e-mail: jordg@cpgen.cpg.com.au
  Transforming the way people work with technology with
  INTEGRITY LEARNING INNOVATION TEAMWORK PERFORMANCE



[-- Attachment #2: Type: text/html, Size: 4452 bytes --]

Re: NFS, Can't get request slot

[Alan Cox]

> Nov 10 04:33:17 spc81 kernel: nfs: server student not responding, still trying
> Nov 10 04:33:17 spc81 kernel: nfs: server student OK

These are all one second or subsecond timeouts

> Nov 10 04:45:56 spc81 kernel: nfs: server student not responding, still trying
> Nov 10 04:46:00 spc81 kernel: nfs: task 20457 can't get a request slot
> Nov 10 04:46:00 spc81 kernel: nfs: server student OK

One four second one here

> We have changed the NIC on this server to 3Com 3c90x for no change in status.
> Clients are using  CNETPro120 200 and Intel eepro100

If samba is ok its not the NIC, it could be congestion on routers or hubs
but you have tried rsize 1024 (and wsize ??)

> Any more help is much appreciated.

You might want to just bump up the NFS timeouts for those warning messages
a couple of seconds. I dont think you have an actual problem. Timeouts are
configurable in mount


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/