Unix Technical Forum

NFS pecularities

This is a discussion on NFS pecularities within the AIX Operating System forums, part of the Unix Operating Systems category; --> Hello, I observed a little odd behavior of the NFS client and server connectivity between AIX boxes (5.3 TL4 ...


Go Back   Unix Technical Forum > Unix Operating Systems > AIX Operating System

FAQ Members List Calendar Search Today's Posts Mark Forums Read
  #1 (permalink)  
Old 01-05-2008, 12:32 PM
rafal polonski
 
Posts: n/a
Default NFS pecularities

Hello,

I observed a little odd behavior of the NFS client and server
connectivity between AIX boxes (5.3 TL4 client and AIX 5.2 ML9 as a server).

There is a firewall between them so I opened these ports:
port 111 UDP (sun rpc)
port 2049 TCP (nfs)
port 32700-32900 TCP(ephemeral ports)
http://www.ncftp.com/ncftpd/doc/misc...ports.html#AIX
# /usr/sbin/no -a | fgrep ephemeral
tcp_ephemeral_high = 65535
tcp_ephemeral_low = 32768
udp_ephemeral_high = 65535
udp_ephemeral_low = 32768


Now when I check from the client resources on the server:
showmount -e server
I get the filesystems and able to mount them. This is a part of the
tcpdump on the server during showmount -e from the client:

07:34:38.798820823 client.798 > server.32783: tcp 0 (DF)
07:34:38.798926395 server.32783 > client.798: tcp 0 (DF)
07:34:38.800673029 client.798 > server.32783: tcp 0 (DF)
07:34:38.800745118 client.798 > server.32783: tcp 44 (DF)
07:34:38.807053003 server.32783 > client.798: tcp 488 (DF)
07:34:38.810154446 client.798 > server.32783: tcp 0 (DF)
07:34:38.810213400 server.32783 > client.798: tcp 0 (DF)
07:34:38.810392712 server.32783 > client.798: tcp 0 (DF)

32783 is pretty logical, it was served by portmapper:
server # rpcinfo -p |grep mount
100005 1 tcp 32783 mountd
100005 2 tcp 32783 mountd
100005 3 tcp 32783 mountd
100005 1 udp 32812 mountd
100005 2 udp 32812 mountd
100005 3 udp 32812 mountd


I mount the filesystems (NFS3, bg,soft,intr). So far so good.

Few hours later, sometimes just few minutes later when I am trying to
again check the resources on the server I do a showmount -e server and
the command hangs for a minute and finally I get RPC timeout:

# showmount -e server
server: RPC: 1832-018 Port mapper failure - RPC: 1832-008 Timed out


this is the tcpdump of this communication. Started on the server.

07:42:21.333809913 client.836 > server.sunrpc: udp 56
07:42:26.331438846 client.836 > server.sunrpc: udp 56
07:42:36.331665184 client.836 > server.sunrpc: udp 56
07:42:56.330956968 client.836 > server.sunrpc: udp 56


Still the mounts created earlier are accessible.
I do a ls on one of the mountpoins.

tcpdump on the server says:
07:59:42.120892807 client.32785 > server.shilp: tcp 0 (DF)
07:59:42.126034715 client.32785 > server.shilp: tcp 0 (DF)
07:59:42.126534273 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.127702191 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.128826386 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.129876331 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.130878705 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.131904956 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.132952712 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.133935908 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.134966085 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.135989063 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.137110763 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.139064942 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.140050363 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.141256566 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.142258576 client.32785 > server.shilp: tcp 136 (DF)
07:59:42.158518620 client.32785 > server.shilp: tcp 0 (DF)
07:59:45.356356575 client.32785 > server.shilp: tcp 140 (DF)
07:59:45.358219786 client.32785 > server.shilp: tcp 0 (DF)
07:59:46.342325451 client.32785 > server.shilp: tcp 140 (DF)
07:59:46.343288988 client.32785 > server.shilp: tcp 160 (DF)
07:59:46.358420845 client.32785 > server.shilp: tcp 0 (DF)

# grep shilp /etc/services
shilp 2049/tcp
shilp 2049/udp


showmount still hangs. I repeat the command and still no luck:
08:00:32.945658156 client.802 > server.sunrpc: udp 56
08:00:37.943803847 client.802 > server.sunrpc: udp 56

What is interesting now local port is changed to 802, previously it was
836.

Again, after some time showmount -e server starts working.

One more tcpdump. When I am unmounting the filesystem (success):
08:11:59.549821319 client.659 > server.sunrpc: udp 56
08:11:59.552718693 server.sunrpc > client.659: udp 28
08:11:59.612331087 client.659 > server.32812: udp 124
08:11:59.675376805 server.32812 > client.659: udp 24




and mounting it again (success):
08:12:52.656996398 client.659 > server.sunrpc: udp 56
08:12:52.657528865 server.sunrpc > client.659: udp 28
08:12:52.662170488 client.659 > server.32783: tcp 0 (DF)
08:12:52.662279072 server.32783 > client.659: tcp 0 (DF)
08:12:52.665904430 client.659 > server.32783: tcp 0 (DF)
08:12:52.665963042 client.659 > server.32783: tcp 44 (DF)
08:12:52.692099534 server.32783 > client.659: tcp 28 (DF)
08:12:52.692938414 client.659 > server.sunrpc: udp 56
08:12:52.693185757 server.sunrpc > client.659: udp 28
08:12:52.696760037 client.702 > server.shilp: tcp 0 (DF)
08:12:52.696826773 server.shilp > client.702: tcp 0 (DF)
08:12:52.698429550 client.702 > server.shilp: tcp 0 (DF)
08:12:52.698489020 client.702 > server.shilp: tcp 44 (DF)
08:12:52.698608790 server.shilp > client.702: tcp 28 (DF)
08:12:52.699680197 client.659 > server.32783: tcp 128 (DF)
08:12:52.699794730 client.702 > server.shilp: tcp 0 (DF)
08:12:52.699845968 server.shilp > client.702: tcp 0 (DF)
08:12:52.699867684 server.shilp > client.702: tcp 0 (DF)
08:12:52.701097633 client.702 > server.shilp: tcp 0 (DF)
08:12:52.839179105 server.32783 > client.659: tcp 0 (DF)
08:12:52.855485048 server.32783 > client.659: tcp 76 (DF)
08:12:52.856605351 client.659 > server.32783: tcp 0 (DF)
08:12:52.856649832 server.32783 > client.659: tcp 0 (DF)
08:12:52.856800205 client.32785 > server.shilp: tcp 136 (DF)
08:12:52.856964885 server.shilp > client.32785: tcp 116
08:12:52.858082591 client.32785 > server.shilp: tcp 136 (DF)
08:12:52.858169327 server.shilp > client.32785: tcp 116 (DF)
08:12:52.859090871 client.32785 > server.shilp: tcp 136 (DF)
08:12:52.859228051 server.shilp > client.32785: tcp 116 (DF)
08:12:52.860135333 client.32785 > server.shilp: tcp 140 (DF)
08:12:52.860223923 server.shilp > client.32785: tcp 124 (DF)
08:12:52.861099931 client.32785 > server.shilp: tcp 136 (DF)
08:12:52.861180653 server.shilp > client.32785: tcp 168 (DF)
08:12:52.866172958 server.32783 > client.659: tcp 0 (DF)
08:12:52.867225813 client.659 > server.32783: tcp 0 (DF)
08:12:52.935294368 client.32785 > server.shilp: tcp 0 (DF)


Any thoughts of this random behavior?

Regards,

rafal
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
Reply


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On
Forum Jump


All times are GMT. The time now is 05:10 PM.


Powered by vBulletin® Version 3.6.5
Copyright ©2000 - 2008, Jelsoft Enterprises Ltd.
Search Engine Optimization by vBSEO 3.2.0
www.UnixAdminTalk.com