marsh@umt.UUCP (Paul Marsh) (01/31/89)
Reply-to: marsh@umt.UUCP (Paul Marsh) We've got a VAX/785, currently running Ultrix 1.2, and we recently tried to bring up the binary release of 3.0. In doing so, I ran into a number of problems. We're back to V1.2 now, so I don't have any way to experiment around for solutions. In the event that we try and bring V3.0 up in the middle of our quarter, I thought that I could minimize some of our down time if I could get some clues to some of my questions. 1) If I select only the mandatory subsets in the "Supported Software Installation" section. The resulting config file contains, among other things, INET, NFS, RPC, and DLI options. We're currently a standalone host, (uucp only connections, later to have DECNET). Do I gain/loose anything, by removing these options and pseudo-devices? From the descriptions in the "System Configuration File Maintenance" manual, these look unnecessary for our configuration, but I wasn't sure if some other option was dependent upon them. 2) Before trying to mount one of my V1.2 partitions, fsck reported a block that was unreadable. (I assume that fsck has changed since my current version doesn't seem to have any problems with it.) I had a backup, so I ran the system disk exerciser on that partition, (dskx -pra0f). While this was running, I was playing with uerf, similar to "uerf -r 300-301 > /tmp/urrf.ss". Sometime later, while these two processes were running, we experienced a "segmentation fault". Which brings me back to fsck. On the /tmp partition, I got fsck errors. PARTIALLY TRUNCATED INODE I=3 (CORRECTED) 2079158789 BAD I =3 UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY Which I did, and ended up with a number of messages. 2079158789 BAD I=3 2079158790 BAD I=3 2079158791 BAD I=3 2079158792 BAD I=3 2079158793 BAD I=3 2079158794 BAD I=3 2079158795 BAD I=3 2079158796 BAD I=3 2079158790 BAD I=3 2079158791 BAD I=3 EXCESSIVE BAD BLKS FOR I=3 CONTINUE? y INCORRECT BLOCK COUNT I=3 (336 should be 226) CORRECT? y I experienced similar errors with I=5. This was followed by these errors for 645 different "I" values. SPARE FIELDS OF INODE CONTAINED VALUES I=xx CONTINUE? y When fsck was done, things were fixed, and the /tmp partition was EMPTY. This case was no big loss, since it was the temporary area. My concern would be if it were my "user" partition, or some other valuable partition. Could this be a problem with V1.2 partitions being mounted under V3.0? Any insight would be helpful. Thanks in advance. -- UUCP : ... ! ucdavis ! umt ! marsh US Mail : Paul Marsh, University of Montana, Computer Center Missoula, MT 59812 phone : (406) 243-5455