Memory leak in cluster NFS
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Warnes, Gregory R gregory_r_warnes at groton.pfizer.comMon Jan 8 09:36:17 PST 2001
- Previous message: Scyld, Athlon, and Asus A7V?
- Next message: job queue/farm out suggestions
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
There is a patch available in the bug reporting system at http://ClusterNFS.sourceforge.net that fixes this problem. I will be releasing a new version of the software shortly, now that I again have a Unix system at my disposal. ClusterNFS was designed for the situation where the nodes have mixed sequential/parallel use. The purpose and advantage of ClusterNFS over "PXE+DHCP+hardlinked passwd/shadow" file configuration comes when you want to maintain the systems you've installed. If you need to update some feature of the configuration on all of the machines, there are 2 problems: 1) it is necessary to connect to each machine and individually make the necessary change and 2) it is difficult to keep track of which machines have been updated. Both of these become even more difficult when there are long running sequential jobs that should not be interrupted. Greg Warnes <greg at warnes.net> -----Original Message----- From: Jeong Hwan Park [mailto:jhpark at brainz.co.kr] Sent: Thursday, December 07, 2000 7:56 PM To: Bjoern Rabenstein; Beowulf Mailinglist Subject: RE: Memory leak in cluster NFS I had a same problem with you. I have changed the Cluster NFS to kernel NFS. I think the ClusterNFS is not great helpful to diskless cluster and it needs a same cost with "PXE+DHCP+hard linked passwd/shadow file configuration" > -----Original Message----- > From: beowulf-admin at beowulf.org > [mailto:beowulf-admin at beowulf.org]On Behalf Of Bjoern Rabenstein > Sent: Thursday, December 07, 2000 11:30 PM > To: Beowulf Mailinglist > Subject: Memory leak in cluster NFS > > > We just have tried ClusterNFS (http://ClusterNFS.sourceforge.net ). > We observe a significant memory leakage. (To prevent crashes, we have to > restart the nfs-server every couple of hours.) > There is an (unresolved) bug report on sourceforge > (http://sourceforge.net/bugs/?func=detailbug&bug_id=122310&group_id=5768 > ). But I am wondering why the memory leak was discovered so late. (The > current release of ClusterNFS is from November last year.) So perhaps the > problem occurs only under certain circumstances (big hard drives, special > file systems...). > Any experiences? > -- > Bjoern Rabenstein * PhD student * Freie Universitaet Berlin * Fachbereich > Biologie/Chemie/Pharmazie * Inst. f. Chemie * Takustr. 6 * D-14195 Berlin > [email] rabe at chemie.fu-berlin.de [phone] +49-30-838-53484 > [WWW] http://lie.chemie.fu-berlin.de/~rabe/ [fax] +49-30-838-53464 > > _______________________________________________ > Beowulf mailing list > Beowulf at beowulf.org > http://www.beowulf.org/mailman/listinfo/beowulf _______________________________________________ Beowulf mailing list Beowulf at beowulf.org http://www.beowulf.org/mailman/listinfo/beowulf
- Previous message: Scyld, Athlon, and Asus A7V?
- Next message: job queue/farm out suggestions
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
