On Tuesday, January 31, 2012 10:49:24 AM Mark Wendt did opine: > On 01/31/2012 09:42 AM, gene heskett wrote: > > On Tuesday, January 31, 2012 09:37:52 AM Mark Wendt did opine: > >> On 01/31/2012 09:02 AM, gene heskett wrote: > >>>> NFS can use up all the CPU cycles trying to get a remount. I've > >>>> seen that quite a few times, especially at boot time, when a > >>>> system is trying to do an NFS mount on another system that's down > >>>> at the time, and it just sits there forever until the other > >>>> machine comes back up. The soft and retrans option allows the > >>>> boot to continue after so many attempts, and then you can fix the > >>>> problem once it's fully booted. > >>>> > >>>> Mark > >>> > >>> But, this is _after_ the other machine had been rebooted for 15-20 > >>> minutes when this one locked up, hence my tendency to think it > >>> wasn't related to any traffic thru NFS at the time. I had just > >>> logged back in over ssh, done a cd to a subdir and an ls. The next > >>> operation and I forget now what it was, was over that ssh link, > >>> might have been an ls -l, completed ok, but then the keyboard and > >>> mouse were gone, I looked over at the gkrellm strip, it was frozen, > >>> I then reached and found the reset button. Had it been synchronous > >>> to the shop machines reboot, that would have been another critter > >>> entirely to this, IMO. > >>> > >>> Cheers, Gene > >> > >> That's what is normally called a stale file handle. The "soft" > >> attribute will help with that also. For whatever reason, when the > >> server side was rebooted, the client did not cleanly dismount, or > >> never dismounted the NFS mount. It's still trying to give you > >> information from the last mount, which doesn't exist anymore, and it > >> keeps trying to process the NFS commands. If you had done a manual > >> "umount" then a manual "mount" of the NFS partition before trying to > >> access it with the "ls" command it probably would have been okay. > >> This is also where the automount facility can come in real handy > >> too. Automount automagically umounts an NFS partition after a > >> certain period of time of non-use. The NFS partition will then be > >> automounted the next time you access it. Check out autofs. There > >> are a lot of configuration settings that will make NFS mounts > >> virtually painless. > >> > >> Mark > > > > autofs is running on both boxes. No clue what the dismount timeout is > > but I would have thought it would have expired. I use it for copying > > stuff back and forth, but anything else is done over an ssh -Y link, > > which should not involve NFS. > > > > Cheers, Gene > > That's really odd. Which machine is the one that's locking up, the > Ubuntu or the pclos? > > Mark
This one, pclos. But historically, it has happened before running mdv or fedora on this box when I was trying to make nfs work. I agree, but see rule #1 ;-) Cheers, Gene -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) My web page: <http://coyoteden.dyndns-free.com:85/gene> "Never underestimate the power of a small tactical nuclear weapon." ------------------------------------------------------------------------------ Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
