Re: FW: RW on init node's hard drive
John Byrne <[email protected]>
| Newsgroups | gmane.linux.cluster.ssic.user |
|---|---|
| Message-ID | <[email protected]> |
As to NFS, are you running nis? Getting nis to work correctly on all nodes is tricky and that could cause problems on nodes other than 1. Also, do have a cvip address set up? What are the export parameters on the NFS server? As to CFS, I know of no reason you shouldn't be about to create directories in a CFS file-system. However, I don't know as I ever used the option "errors=remount-ro". Could you try removing this mount option? (Depending on your distro, you may need to rebuild your initrd to make sure it goes away entirely.) Thanks, John Byrne Walker, Bruce J (HP-Labs) wrote: > > > -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of J. Miribel > Sent: Wednesday, November 29, 2006 10:51 AM > To: Daniel 'Fremen' Llewellyn > Cc: [email protected] > Subject: Re: [SSI-users] RW on init node's hard drive > > > > Daniel 'Fremen' Llewellyn a écrit : >> On 29/11/06, J. Miribel <[email protected]> wrote: >>> Hi everyone, >>> >>> I'm running openssi 1.9.1 over debian and I'm facing a little problem >>> I can't fix for now: >>> I can't get write privileges on my nfs mount on any of my nodes >>> except the init node: >>> maitre1:~# onnode 18 mkdir /home/huhu >>> mkdir: Ne peut créer le répertoire `/home/huhu': Permission non >>> accordée maitre1:~# >>> >>> However, this works on node 1: >>> maitre1:~# onnode 1 mkdir /home/huhu >>> maitre1:~# >>> >>> Here is my fstab: >>> proc /proc proc defaults,node=* 0 0 >>> /dev/sda1 / ext3 defaults,errors=remount-ro,node=1 >>> 0 1 >>> /dev/sda5 none swap sw,node=1 0 0 >>> 192.168.0.2:/usr/local/games /usr/local/games nfs >>> node=*,soft,timeo=5,intr,rsize=8192,wsize=8192 0 0 >>> 192.168.0.2:/home/maitre1 /home nfs >>> node=*,soft,timeo=5,intr,rsize=8192,wsize=8192 0 0 >> wouldn't it be easier to utilise oSSI's cluster filesystem to access >> everything rather than adding another layer of complexity by using nfs >> over cfs over 'your hard drive'? ie, if cfs is already being used, >> then why not stay with that instead of REmounting your /home dir with >> nfs pointing to the /home/maitre1 dir which is being managed by CFS? >> >> I may be wrong in my assumption that you are running with cfs, but I >> can't see why cfs would be dropped from 1.9 when it is available in >> the stable versions >> > Hi > > In fact I don't use cfs because my nfs server ain't on my cluster... > Problem is the same if I try to create /test which would be on init node > #1 on drive/dev/sda1, so this problem doesn't seem to be due to NFS but to the cluster itself... > > Is that normal ? > > Best regards > > Regards > > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > Ssic-linux-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/ssic-linux-users > ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV