[Top][All Lists]

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


From: Steve Ramage
Subject: CVS and NFS
Date: Fri, 9 Mar 2007 14:57:29 -0800

The long question:
I was investigating the best Windows CVS client for us to use, and noticed that every client seemed to have a problem dealing with Network Drives. Upon further investigation I noticed that there have been alot of reports about CVS and NFS in general not being safe however the main threads dedicated to discussing this issue seem to be unable to say what exactly is wrong, and what the actual level of risk is []. They also were dated in 2004 and so I wasn't sure if it was topical or not.
So just to be clear our current setup is this:
Dev-Box-A -----
               |----(/home NFS mount)----                       
Dev-Box-B -----|                         |---NAS Device.
               |----(/cvsroot NFS mount)-
Dev-Box-C -----
So all boxes have two NFS mounts, that mount to different shares on the NAS devise. On each Dev Box we basically check out from cvsroot and check into a folder in our home directory. It seems that the /cvsroot mount is hard mounted? Where as the /home mount is soft mounted? Both seem to have interrupt enabled, and both seem to be using TCP and NFS version 3.All three boxes, are Solaris (9 and 10) and we have one sparc and two intel. Its possible that any developer could be using any box to develop and commit to any project, concurrently.
What is the risk of corruption to the repository, and or local working copy in the above setup? Some postings have suggested that NFS3 with hard mounting, and interrupts enabled should be fine. Others suggest that at any time the repository may be corrupted and we wouldn't notice. Should we try to move both the repository and the development enviroment away from remote file systems? Is there only a risk to the repository, and not to the local copies since they will never be touched concurrently on two machines at the same time?
The short question:
I guess the short question to this post is the discussion in still mostly relevant to modern day NFS implementations and NAS storage devices, and do we risk NFS corruption.
Steve Ramage

reply via email to

[Prev in Thread] Current Thread [Next in Thread]