home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.nfs
- Path: sparky!uunet!usc!sol.ctr.columbia.edu!ira.uka.de!math.fu-berlin.de!informatik.tu-muenchen.de!LRZnews!k2
- From: k2@bl.physik.tu-muenchen.de (Klaus Steinberger)
- Subject: Re: NFS corruption discovery
- Message-ID: <k2.714296050@woodstock>
- Sender: news@news.lrz-muenchen.de (Mr. News)
- Organization: Leibniz-Rechenzentrum, Muenchen (Germany)
- References: <1992Aug19.225010.18306@den.mmc.com> <1992Aug20.042224.11176@decuac.dec.com>
- Date: Thu, 20 Aug 1992 07:34:10 GMT
- Lines: 26
-
- mjr@hussar.dco.dec.com (Marcus J. Ranum) writes:
-
- >>I've also looked at lan traffic and I've discovered the UDP checksum is off
- >>for NFS packets. Unfortunately the net isn't reliable enough [...]
-
- >turned off, so they can benchmark a little bit faster. Others sell
- >versions of NFS that do asynchronous writes. Some of them will tell
- >you of the risks, and others won't.
- The interesting thing is, that the vendor(s) who run with UDP checksum off
- don't tell you that, but the vendors who offer asynchronous NFS are telling you
- what you are doing. And the hazzles with UDP checksumming off are really
- big compared with the mostly nonexistant hazzles of asynchronous NFS.
-
- > Digital's ULTRIX runs with UDP checksumming on *ALWAYS*
- >and with synchronous NFS *always*.
- I agree that UDP checksumming should be always on, but asynchronous NFS
- should really be a option, as you win really much performance.
- Putting your server onto a UPS is always a better solution!
-
- Sincerely,
- Klaus Steinberger
- --
- Klaus Steinberger Beschleunigerlabor der TU und LMU Muenchen
- Phone: (+49 89)3209 4287 Hochschulgelaende
- FAX: (+49 89)3209 4280 D-8046 Garching, Germany
- Internet: Klaus.Steinberger@Physik.Uni-Muenchen.DE
-