home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!destroyer!cs.ubc.ca!l4x092
- From: l4x092@rick.cs.ubc.ca (Anthon Su Tong Pang)
- Newsgroups: comp.sys.sun.admin
- Subject: Phantom FS errors? (REPOST)
- Date: 26 Jan 1993 00:56:17 GMT
- Organization: Computer Science, University of B.C., Vancouver, B.C., Canada
- Lines: 32
- Message-ID: <1k227hINNh5o@cs.ubc.ca>
- NNTP-Posting-Host: keats.rick.cs.ubc.ca
-
- For the past month (since adding a single NFS client--Sun 3/75--to our
- Sun 3/50), we've been plagued with phantom FS errors. Everything runs
- fine for a while, but then jobs (new or already running) start to die
- like flies complaining about filesystem related problems.
-
- For example, when compiling gcc 2.3.3 on the client, 'make' eventually
- dies complaining about a bad object file. We 'sync' up and run 'fsck'
- on /usr on the server. Many errors are reported (partially allocated
- inodes, truncated inodes, bad mode, etc). Since /usr has been mounted
- read-only by the server and client through out, we decide NOT to let
- 'fsck' repair the filesystem. Instead we 'reboot' the server.
-
- We've installed the UFS and NFS jumbo patches for 4.1.1. But everything
- runs fine if we only use the server...when we run these (disk intensive)
- "torture" tests on the NFS client, FS errors are reported.
-
- Does anyone know of a possible cause for these 'now-they're-here'
- and 'now-they're-not' filesystem errors? And how do we get rid of
- them?
-
- Our suspicion is that the super block in memory is getting munged (since
- the problems disappear after a reboot).
-
- Oh, and while I have your attention...'reboot' always halts with a
- BUS error on our 3/50. Is there a fix for this?
-
- Many thanks in advance.
-
- --
- #include <disclaimer.h> /* CSSS -- CS Dept. Rep. */
- #include <amiga.h> /* my machine is better than your machine */
- #include <ocp.h> /* Amiga Developer */
-