rog@speech.kth.se (Roger Lindell) (06/06/91)
Hello, we have recently acquired five 9000/425s machines and have installed one Micropolis 1598-15 1.2 Gb disk in each of them. This configuration works fine except for a few cases, and these are: 1. If I try to duplicate a directory tree on the same disk as the original directory tree using the command 'cp -opPrs old_dir new_dir' the machine may crash, and if it does crash it takes about 20 minutes for it to get down to the mnemonic debugger level. If I have the machine in service mode I get an error code that either is 80380018 [no resources available (OS/SCSI manager)] or 80080003 [disk controller time-out (OS/disk manager)]. 2. If there is heavy use on the disk an error might get logged that says, if you use lsyserr, "disk driver logic error (OS/disk manager)" and also "disk error recovered by some device action (OS/Recovered Disk Error Manager)". This has happened on 3 out of 5 machines which makes it, in my opinion, unlikely that there is a hardware error. BTW this error does not crash the machine. 3. We added a WangDAT to one of our 9000/425s and used SCSI-device 1 and are therefore able to use rbak/wbak with device m to access the DAT. The strange thing here is that if I make a wbak to the DAT with files from the local disk the machine is liable to hang and I get the error code 80080003 [disk controller time-out (OS/disk manager)] and the same thing happens if I rbak some files that were taken from one of our other disks to the local disk. This error does not happen if I do rbak/wbak to/from the other disks in our network, or if make a large wbak that includes all our disks, including the disk on the machine that has the DAT. Has anybody out there experienced the same thing or something like it, and is there a solution to our problems? Many thanks in advance, -- Roger Lindell rog@speech.kth.se Phone: +46 8 790 75 73 Fax: +46 8 790 78 54 Dept. of Speech Communication and Music Acoustics Royal Institute of Technology Sweden
bep@quintro.uucp (Bryan Province) (06/07/91)
In article <rog.676192085@system> rog@speech.kth.se (Roger Lindell) writes: > >we have recently acquired five 9000/425s machines and have installed one >Micropolis 1598-15 1.2 Gb disk in each of them. This configuration works >fine except for a few cases, and these are: > >1. If I try to duplicate a directory tree on the same disk as the original >directory tree using the command 'cp -opPrs old_dir new_dir' the machine >may crash, and if it does crash it takes about 20 minutes for it to get down >to the mnemonic debugger level. If I have the machine in service mode I get >an error code that either is 80380018 [no resources available (OS/SCSI >manager)] or 80080003 [disk controller time-out (OS/disk manager)]. > >2. If there is heavy use on the disk an error might get logged that says, >if you use lsyserr, "disk driver logic error (OS/disk manager)" and also >"disk error recovered by some device action (OS/Recovered Disk Error >Manager)". This has happened on 3 out of 5 machines which makes it, in my >opinion, unlikely that there is a hardware error. BTW this error does >not crash the machine. We had similar kinds of errors on an external SCSI disk on our 400t. It was a third party drive but I called Apollo to have them explain the errors anyway. They said that it corresponds to a bad spot on the disk. So I used invol and marked the address given, by lsyserr, as bad. I was under the mispresumption that the SCSI disks took care of badspots by themselves. Well they do at invol time but any further errors have to be logged in the badspot list. We haven't had any problems since. If your lsyserr gives you any address (especially repeating ones) try logging them in the badspot list, run a long salvol, and try running for a while to see what happens. Also check that the end of your SCSI chain is terminated correctly. Make sure that everything is "single ended" type of SCSI. Also it seems, at least on our 400t, that no terminator is needed internally nor on the internal disk; just on the external port or on any external device(s). (BTW, the drive and third party vendor are the ones I was complaining about several months ago, so everything is fine now.) -- --=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-- Bryan Province -Glenayre Corp., Quincy, IL- quintro!bep@lll-winken.llnl.gov "I tried putting instant coffee in the microwave, I almost went back in time." - Steven Wright