Problem with NFS from RedHat4 to Solaris

Problem with NFS from RedHat4 to Solaris

[Shane Presley]

Hello,

We have a RedHat3 server, with some NFS mounts to Solaris9 servers.
That works nicely.  We also run samba on top of that, for Windows
access to the NFS shares.  Works fine.

However, on a RedHat4 system, using the exact same mount, we
experience odd behavior.  Through the command line, CAT'ing a file
will show the file, but fail to return you back to the command line
immediately.  It's as if it's waiting for the EOF characters?
Eventually (2-5 minutes) it returns you to the command prompt.

My mount looks like

mount -t nfs  -o soft,intr,rsize=1024,wsize=1024,nosuid,udp
mysolariserver:/opt/path /misc/path

Same results autofs vs command line mount.  I have tried playing with
the rsize/wsize, seems to have no effect.  Also, I have tried removing
the udp option, same result.

RedHat3 system that works...
autofs-4.1.3-186
nfs-utils-1.0.6-44EL

RedHat4 system that hangs...
autofs-4.1.3-187
nfs-utils-lib-1.0.6-3
nfs-utils-1.0.6-70.EL4


Looking at a snoop of the traffic between the servers shows that the
RedHat4 system seems to get hung up in a loop with about two minutes
of this:

myredhat4server -> mysolaris9server          PORTMAP C GETPORT
prog=100021 (NLM) vers=4 proto=UDP
         mysolaris9server -> myredhat4server PORTMAP R GETPORT port=0
myredhat4server -> cs8          PORTMAP C GETPORT prog=100021 (NLM)
vers=4 proto=UDP
         mysolaris9server -> myredhat4server PORTMAP R GETPORT port=0

Where as on the RedHat3 snoop, I never see any of that PORTMAP traffic?

Any idea where I should look?

Thanks
Shane

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Problem with NFS from RedHat4 to Solaris

[Steve Dickson]

Shane Presley wrote:
> However, on a RedHat4 system, using the exact same mount, we
> experience odd behavior.  Through the command line, CAT'ing a file
> will show the file, but fail to return you back to the command line
> immediately.  It's as if it's waiting for the EOF characters?
> Eventually (2-5 minutes) it returns you to the command prompt.
> 
> My mount looks like
> 
> mount -t nfs  -o soft,intr,rsize=1024,wsize=1024,nosuid,udp
> mysolariserver:/opt/path /misc/path
Do it work with TCP mounts?

> 
> 
> Looking at a snoop of the traffic between the servers shows that the
> RedHat4 system seems to get hung up in a loop with about two minutes
> of this:
Is SELinux enabled?


> Any idea where I should look?
Is there anything error messages in /var/log/message?

steved.

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Problem with NFS from RedHat4 to Solaris

[Shane Presley]

On 10/24/06, Steve Dickson <SteveD@redhat.com> wrote:
> Do it work with TCP mounts?
tcp NFS behaves the same.  I prefer UDP to match the RHEL3 behavior,
but I did try TCP, and had the same problem.


> Is SELinux enabled?
Nope, no SELinux or firewall.


> Is there anything error messages in /var/log/message?
Not a thing.  I tried turning on debug level SAMBA logging, but can't
find anything that jumps out at me during the error.  And of course
this happens even without SAMBA running on top.


I'm wondering if this could be NFS version related..or something to do
with the portmapper service?  I find it odd that when I do a tcpdump
from the RHEL3 box, I never see that portmapper " PORTMAP C GETPORT"
traffic.  Just all UDP NFS packets.

Thanks,
Shane

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Problem with NFS from RedHat4 to Solaris

[Steve Dickson]

Shane Presley wrote:
> 
> Looking at a snoop of the traffic between the servers shows that the
> RedHat4 system seems to get hung up in a loop with about two minutes
> of this:
> 
> myredhat4server -> mysolaris9server          PORTMAP C GETPORT
> prog=100021 (NLM) vers=4 proto=UDP
>          mysolaris9server -> myredhat4server PORTMAP R GETPORT port=0
ah... port=0 means it can't get server's port number for lockd...
make sure the lockd on the server is running...

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Problem with NFS from RedHat4 to Solaris

[Shane Presley]

On 10/25/06, Steve Dickson <SteveD@redhat.com> wrote:
> ah... port=0 means it can't get server's port number for lockd...
> make sure the lockd on the server is running...
>


Very interesting...and you're correct....this mount is on my Solaris servers
rpcinfo -p solarisserver
   program vers proto   port
    100000    4   tcp    111  portmapper
    100000    3   tcp    111  portmapper
    100000    2   tcp    111  portmapper
    100000    4   udp    111  portmapper
    100000    3   udp    111  portmapper
    100000    2   udp    111  portmapper
    100005    1   udp  35185  mountd
    100005    2   udp  35185  mountd
    100005    3   udp  35185  mountd
    100005    1   tcp  54469  mountd
    100005    2   tcp  54469  mountd
    100005    3   tcp  54469  mountd
    100003    2   udp   2049  nfs
    100003    3   udp   2049  nfs
    100227    2   udp   2049  nfs_acl
    100227    3   udp   2049  nfs_acl
    100003    2   tcp   2049  nfs
    100003    3   tcp   2049  nfs
    100227    2   tcp   2049  nfs_acl
    100227    3   tcp   2049  nfs_acl


No nlockmgr.  But for some reason when my RHEL3 system mounts this
server, it still works.  Anyway, now I have to dig into my Solaris box
to see why nlockmgr isn't working.

But any idea why RHEL3 would still work fine?

Thanks
Shane

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs