NFS: When one server goes down, entire network slows down to a crawl.

NFS: When one server goes down, entire network slows down to a crawl.

Post by Gus Protopapa » Wed, 01 Sep 1999 04:00:00



I have this NFS problem where there is one server in the states connected
via NFS to my computer here in Canada.  When the US computer goes down, my
computer in Canada is pretty much locked.  I can't do anything.

Is this some sort of necessary evil I will have to put up with, or have I
configured the NFS wrong?  By the way, things are peachy when they are both
up and running... works really nice.

Thanks for any assistance you may be able to provide,

Gus

 
 
 

NFS: When one server goes down, entire network slows down to a crawl.

Post by Marvin W. Goodma » Wed, 01 Sep 1999 04:00:00




Quote:> I have this NFS problem where there is one server in the states
connected
> via NFS to my computer here in Canada.  When the US computer goes
down, my
> computer in Canada is pretty much locked.  I can't do anything.

> Is this some sort of necessary evil I will have to put up with, or
have I
> configured the NFS wrong?  By the way, things are peachy when they are
both
> up and running... works really nice.

Gus,

I think that more information is required.  What is the nature
of the relationship between the Canadian system and the stuff
on the directory exported from the U.S. system?  In other words,
what's running on the Canadian computer that would suddenly not
be able to run if the mount was taken away?  Is this an automount
directory?

-marv

Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

 
 
 

NFS: When one server goes down, entire network slows down to a crawl.

Post by HB mp » Wed, 01 Sep 1999 04:00:00


NFS, in its nature, is pretty much a bear if the connection is lost between the
two.  I.E.     the server is exporting this filesystem jumping up and down
saying "I have your filesystem"  and when the client boots and sees that the
server is offering, it grabs hold of the servers filesystem(by mounting the
filesystem).  If either of the links brake without unmounting the NFS'd
filesystem, the pipe is hung and so are the daemons on both sides (unless the
other server is down therefore no daemons running to be hung).  When the server
comes back up, you may be able to reconnect that NFS filesystem  automatically.

Sorry for being so colorful, but it has been a long day and I needed the stress
relief.

:-)

Statements and suggestions are made in my humble opinion to offer some help.
Enlighten me with more knowledge and let me continue to grow.
IBM Certified Specialist and HP Certified IT Professional.

 
 
 

NFS: When one server goes down, entire network slows down to a crawl.

Post by Buddika Aluwihar » Thu, 02 Sep 1999 04:00:00


Try mounting the filesystem with the option 'soft', then if there is no
response from the server it will generate an error. The default for NFS
option  is 'hard' which means the system will retry until a response is
received from the NFS Server. These are NFS specific options.
Eg. mount -o soft host:<filesystem>

Buddika


>I have this NFS problem where there is one server in the states connected
>via NFS to my computer here in Canada.  When the US computer goes down, my
>computer in Canada is pretty much locked.  I can't do anything.

>Is this some sort of necessary evil I will have to put up with, or have I
>configured the NFS wrong?  By the way, things are peachy when they are both
>up and running... works really nice.

>Thanks for any assistance you may be able to provide,

>Gus


 
 
 

NFS: When one server goes down, entire network slows down to a crawl.

Post by Geral » Thu, 02 Sep 1999 04:00:00


Are you running NIS with the Master server being the one in the US?
This could also cause similar "lockups" if you lose the Master and
there are no Slave Server(s) for you to bind to. Otherwise, you might
want to look into the "intr" option in addition to "soft" as someone
else has mentioned.

> I have this NFS problem where there is one server in the states connected
> via NFS to my computer here in Canada.  When the US computer goes down, my
> computer in Canada is pretty much locked.  I can't do anything.

> Is this some sort of necessary evil I will have to put up with, or have I
> configured the NFS wrong?  By the way, things are peachy when they are both
> up and running... works really nice.

> Thanks for any assistance you may be able to provide,

> Gus


 
 
 

NFS: When one server goes down, entire network slows down to a crawl.

Post by Jennie Haywoo » Fri, 03 Sep 1999 04:00:00


Uhh....not quite.  NFS is stateless.  We DO NOT advertise that a file system is
exported.  You can enquire from a client if we have filesystems exported
(showmount).  What usually makes a client hang is a high level lock being taken and
we spin on it waiting for something to happen, and the network goes down or a
router blows up, etherswitch is being over run, the server went down......
for a hard mount we will sit and wait for ever.....the loop that makes those calls
looks like:
     for(;;){
       stuff
      }
We won't leave it unless there is some catastrophic error code returned.

NFS is not a 500 pound gorilla.......DFS is.  ;-)

What you've described is more how NIS works.......it does advertise......

NIS will get you in a headlock......and NOT let go.


> NFS, in its nature, is pretty much a bear if the connection is lost between the
> two.  I.E.     the server is exporting this filesystem jumping up and down
> saying "I have your filesystem"  and when the client boots and sees that the
> server is offering, it grabs hold of the servers filesystem(by mounting the
> filesystem).  If either of the links brake without unmounting the NFS'd
> filesystem, the pipe is hung and so are the daemons on both sides (unless the
> other server is down therefore no daemons running to be hung).  When the server
> comes back up, you may be able to reconnect that NFS filesystem  automatically.

> Sorry for being so colorful, but it has been a long day and I needed the stress
> relief.

> :-)

> Statements and suggestions are made in my humble opinion to offer some help.
> Enlighten me with more knowledge and let me continue to grow.
> IBM Certified Specialist and HP Certified IT Professional.

 
 
 

1. Interfaces don't go down when network is physically down

On FreeBSD 3.0-RELEASE, when you physically break an ethernet (eg, unplug
the cable) the host still thinks the interface is up and can still
ping it.  Is there any way to avoid that scenario - ie, when you break
an interface, have it be really broken, down, unpingable?

The reason I ask is that a physically down ethernet network interface
does not appear to be visible FreeBSD (or to gated) which means that
OSPF really doesn't work properly.

For example, two machines, three networks:

            +------+            +------+
----net1----| box1 |----net2----| box2 |----net3----
            +------+            +------+

If box1 is announcing net1 to box2 via OSPF, and I physically take
down the net1 interface on box1, box1's gated never knows it, keeps
announcing the route to box2.  box2 can still reach the net1 interface on
box1, even though net1 should really be unreachable via box1.

Any work-around for this?

[FYI, this is not really a FreeBSD-specific problem - Linux and Solaris
also appear to fail this test.]

--Paul

2. access to 2nd ide hd. HELP!

3. How to setup a secondary mail server in case the first one goes down?

4. segmentation fault in f2c-generated prog, URGENT

5. Diald + SLIP slows down to a crawl

6. Korn shell "cascading" xtrace?

7. FTP slows down to a crawl

8. Touch tone modem access

9. Kde 3 slows down to a crawl

10. Added 4gb hard disk and system slows down to a crawl

11. HUB slowing down network with only one user

12. NFS client hangs when the server goes down

13. Need Help with dynamic unmounting when server goes down: Sun NFS