Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!mit-eddie!think!ames!ucbcad!ucbvax!tg.UUCP!scott From: scott@tg.UUCP.UUCP Newsgroups: mod.computers.vax Subject: Submission for mod-computers-vax Message-ID: <8703212338.AA15643@tg.UUCP> Date: Sat, 21-Mar-87 18:38:29 EST Article-I.D.: tg.8703212338.AA15643 Posted: Sat Mar 21 18:38:29 1987 Date-Received: Sun, 22-Mar-87 16:42:06 EST Sender: daemon@ucbvax.BERKELEY.EDU Distribution: world Organization: The ARPA Internet Lines: 56 Approved: info-vax@sri-kl.arpa Path: tg!scott From: scott@tg.UUCP (Scott Barman) Newsgroups: comp.sys.dec,comp.unix.questions,comp.unix.wizards,comp.bugs.4bsd,mod.computers.vax Subject: dump hangs while dumping Eagles Keywords: dump, Eagles, HELP! Message-ID: <160@tg.UUCP> Date: 21 Mar 87 23:38:28 GMT Organization: Townsend Greenspan & Co., Inc. New York, NY Lines: 46 I hope somebody can give me a pointer as to where to find/fix this problem! First the system: DEC VAX 11/780 4 Mb of 'C'-type memory 2 DEC RM05 1 DEC TU77 2 Fujitsu Eagle Drives (emulating expanded RM02s) 1 Emulex SC7002 Controller 1 Emulex V-Master Massbus Extender Running 4.1bsd The problem: The dump program hangs after encountering an ECC error on the Eagle drives. Another problem is that the error cannot be consistantly recreated. Recreate by: We use dump(8) level 0 as a full backup of a partition (by never specifying 'u' on the command line). Doing a full dump on a filesystem that has some bad sectors will cause dump to hang occasionally. These sectors are not in the bad sector file because the Emulex maintenance programs does not allow for the entry of new sectors not found during formatting. Also, the Emulex version of "READALL" does not report the sector numbers found to be bad (not critical since I can figure out the number from the system errors and where the partition begins). These sectors have been "marked" with badsect(8). Additional notes: The backups are done from single user mode. To get the single user, the system is rebooted instead of killing processes. These errors do not occur on the RM05s nor are there any errors on the tape drive. To reset from this error, we have to \ from dump and reboot the system (if we do not reboot, the error will occur again). The particular partition this happens on has two sectors that generate ECC errors and the problem only happens (when it does) on the first. HELP! If you can help, please e-mail suggestions to me. Any information I can get will be appreciated. Also, please don't tell me I should upgrade to 4.3bsd. I already know this and have been trying to convince others of this as well. Thanks! Scott Barman {philabs, pyrnj}!tg!scott