Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: NFS mounting, Hard -vs- automount



 ref wrote: 
> Hi Folks, 
> 
> I use a lot of NFS mounts between my data servers and the front end 
> boxes that use the data to delivery web content. However, I am running 
> into problems when the NFS mounts seem to just .. ' go away '. These are 
> CentOS boxes, and not all mounts to the machine failed, actually only 
> one of five. However, as they fail silently, it sucks... 

You mean the NFS server failed?  Or the server was still alive but suddenly 
the client would hang when accessing?  Or the server was alive but the client 
just magically unmounted the filesystem? 

> I tried to use Automount last year to handle the mounting, but had 
> problems with that too when machines got rebooted. 

This might help: 
  # chkconfig --level 345 autofs on 

> so .. my question is : what are the best ways to keep NFS mounts alive, 
> and monitored ? are hard mounts plus a monitor better in the long run 
> than automount, Are my experiences with automount abnormal, and I should 
> go back with it ? whats the general view here ... 

I've never seen automount cause issues beyond what bare NFS has, especially if 
you avoid indirect maps (ie that involve an '&'). 

I find that using the 'intr' option if you're using 'hard' is really useful. 
It sucks when the NFS server dies, you try and test something on the client, 
and it just hangs forever, Ctrl-C won't do anything.  'intr' will help with 
the 'ctrl-c' issue. 

> I appreciate any responses ... including ones that may say NFS is a bad 
> solution and I shoud use XYZ solution. I am open to anything at this 
> point. 
> All my Servers are CentOS5.1, using NFS4... 

It's hard to say whether NFS is good for you or not, because there are a lot 
of details missing: 
   - is there a reason each front-end box can't have a local copy of the data? 
     - do you have (or want to have) failover NFS servers? 
   - is it necessary to use the 'hard' mount option (as opposed to 'soft')? 
   - if the answer to the first two is 'yes', would GFS 
(http://en.wikipedia.org/wiki/Global_File_System) help?  (note: using 
CentOS-5.x probably helps here, since RHEL5 has GFS built in I think). 

Another random note, last time I used NFS4 on linux it was pretty flaky, and 
kept crashing my server.  NFS3 has rock solid for me for ages. 

HTH, 
Matt 

-- 
This message has been scanned for viruses and 
dangerous content by MailScanner, and is 
believed to be clean. 

_______________________________________________ 
Discuss mailing list 
[hidden email] 
http://lists.blu.org/mailman/listinfo/discuss
 


BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org