NFS mounting

[Mike Burk]

How can I set up my NFS clients to continue to boot if the NFS server is not alive. I have set the mounts as "soft" and it will eventually finish booting but it takes 20 minutes. If the server is alive it only takes 3 minutes to boot? I would like the client to try for 30 seconds and then go on.

Thanks,

Mike

[Mark Vollmers]

Mike-

I've wondered about the same thing, but I havn't really had time to play with it. I suppose that in the rc script where it mounts, you could add your own conditions, something that would be like

if successful contact with server then run mount command else skip

I'm sure that someone out there probably has it written already. Myself, I'm not sure how to do the check for positive connection. step in the right direction, I guess.

Mark

[Pete Randall]

Mike,

I used to have this problem back in the 10.20 days. I now use hard mounts in 11.0 and 11.i and don't have the problem. Perhaps you need to get current on NFS and/or patches.

Good luck,
Pete

[Sanjay_6]

Hi Mike,

Try the NFS mount options with the mount command or in /etc/fstab,

http://docs.hp.com/cgi-bin/fsearch/framedisplay?top=/hpux/onlinedocs/B1031-90048/B1031-90048_top.html&con=/hpux/onlinedocs/B1031-90048/00/00/9-con.html&toc=/hpux/onlinedocs/B1031-90048/00/00/9-toc.html&searchterms=mount%20timeout&queryid=20020522-081958

Also Take a look at this thread,

http://us-support3.external.hp.com/cki/bin/doc.pl/sid=c41f4e7f0a403634b7/screen=ckiDisplayDocument?docId=200000006441210

The Doc id is :CDS9209231413

Hope this helps.

Regds

[S.K. Chan]

I think you can use the "timeo" option in your fstab file. A default for soft NFS TCP mount is ..
timeo=600
i.e 60 secs. However I have not try this before, so couldn't comment much.

[Sachin Patel]

Hi Mike,
With timeo you can use retrans and retry for faster recovery.

if you only have timeo then
timeo=600 = 60 seconds then if it can't mount then it will try 120 second then 240second then 480second then it will timeout. that is because value of retrans is 4 so it is retransmitting 4 times
Also you can specity retry option the default is 1.

Sachin

[BSSG]

Try putting a "retry={some number} in your /etc/auto_master file. It should tell it to stop after so many minutes of trying. Couldn't hurt.

[sven verhaegen]

depending on what causes you issue there are limited ways to get around it , if the server is just slow to reply , you could play with the retrans values , if however the server is down the simplest is to use soft mounting option it will timeout on a mount request after a while (limited time), for sure DO NOT USE 'timeo' in udp environments , this parameter in a UDP environment isn't advised to be used at all and I can tell becuase I've seen some problems with it before , you can use it effectively in a NFS TCP environment as there it's functionality does become effective