holtz@cascade.carleton.CDN (Neal Holtz) (03/22/90)
I have a CDC Wren V SCSI disk (702 MB unformatted) disk on loan from a local company that doesn't know Apollos and is interested to see if it will work. We are trying to add it as a second disk (external) to a DN2500. The DN2500 seems to largely recognize it (cntrlr 5, lu 0). DEX can determine the Manufacturer, ROM version, etc. (except it doesn't seem to get the block count right, but then its not formatted yet). It passes the self tests (except for volume label). I can't get it formatted. After almost 2 hours, invol dies with disk-controller timeout (transcript appended to this message). We have not yet: 1. checked the drive to see if we can format it on another machine, or 2. talked to Apollo service to see if this should work, or 3. tried a shorter cable (currently looks like about 3m). In the meantime, does anyone know any reasons why this shouldn't work? ===================================================== Tried many variations, all with the same result. Under MD: > RE > DI SD5:0 > DI W > BOOT Then under AEGIS, we use invol and choose option 7 -f, then 1: $ invol invol (initialize_volume), revision 10.2, September 5, 1989 3:00:18 pm Options are: 0 - EXIT. 1 [-fnb5uom] - initialize virgin physical volume. 2 [-fnb5u] - add a logical volume. 3 [-fnb5] - re-initialize an existing logical volume. The following flags apply to options 1 thru 3, as indicated: f: don't re-format disk u: don't prompt user - use defaults o: make sr9 format disk n: make non-bootable volume b: apply bsd unix acls 5: apply sys5 unix acls m: build a multi-disk (e.g., striped) group 4 - delete a logical volume. 5 - list logical volumes. 6 [-e] - list badspots on disk or volume ... -e: list in decimal. 7 [-f] - initialize physical badspot list. 8 - create or modify an os paging file. 9 - add to existing badspot list. 10 - display/change sector interleave factor. 11 - remove from existing badspot list Option: 7 -f WARNING: this option will re-initialize (i.g., destroy) the physical badspot list on the disk. Normally, it only needs to be performed when a disk is first received from the factory. If you want to change your mind, type 'q' to the following prompt. Select disk: [w=Winch|s=Storage mod|f=Floppy|q=Quit][ctrl#:][unit#] w5:0 Enter badspots to be ADDED between physical disk addresses 0 and 8F4D3 (hex) Badspots must be input in (hex) physical disk address form. (The manufacturer supplied badspots are handled by the disk controller; there is no need to manually add the badspots listed on the drive.) Terminate badspot entry with a blank line. : Is the badspot information you entered correct? y Done. Anything more to do? y Option: 1 Select disk: [w=Winch|s=Storage mod|f=Floppy|q=Quit][ctrl#:][unit#] w5:0 Physical volume name: holtz Milestones are not available during format of a logically addressed disk. Formatting. Unrecov I/O error during format. abort - disk controller time-out (OS/disk manager) Run aborted.
krowitz%richter@UMIX.CC.UMICH.EDU (David Krowitz) (03/23/90)
We have a Sony SMO-503 erasable-optical disk (magneto-optical disk) which we bought from Workstation Solutions. We found a number of things with the DN2500's SCSI support: 1) Our original cable was an 8-foot cable made with twisted pair wire (round cable). This proved to be unreliable. Workstation Solutions shipped us a 3-foot ribbon cable which has worked well ever since it was installed. They are apparently evaluating several other cables as this time. 2) Our original boot proms were version 1.8 (pop the lid on the DN2500 and look for a socketed chip with a paper tag on it which says something like "FRODO 1.xxx" on it -- this is your boot prom chip). We found that even though we could format our drive we could not boot from it, and occassionally the system seemed to timeout on the drive -- ie. we'd have the drive mounted and be accessing some files on it and all of a sudden we'd get a "file not found" or "directory not found in pathname" error, and when we retried the operation the file was ok. The Sony drive is a fairly slow disk (90 msec average access), so we assumed that the timeouts were a problem with the drive's speed, but it may be a generic problem with external disks. Workstation Solutions advised us that the rev 1.9 boot proms would fix the problem of booting from the optical disk and might also fix the apparent drive timeouts. Our field service reps obtained some new CPU boards for us (apparently the proms are not available as a seperate part yet), and both problems have gone away. 3) With either set of boot proms, we found that even though the DN2500's self test (invoked by the memonic debugger's TE command) would see the disk on the SCSI port, we had to issue the MD command sequence: > RE > DI SD0:0 (our drive is set to device 0 on controller 0) > RE > DI N (we boot diskless from the network, you could give a DI W, DI E, DI F, or DI C to boot from a device other than the external disk) > EX DOMAIN_OS If we did not issue the DI SD command, the machine would boot, but /com/invol, /com/salvol, /com/mtvol, etc. will not recognize the disk. With the rev. 1.8 boot proms, we found that we had to run invol using option 7 first (to create a bad spot list), and then option 1 (to initialize the volume). With the rev 1.9 boot proms, we have to use option 7 -f and 1 -f (ie. do not attempt to reformat) or we will get a status code 8002a (illegal command for device, or something like that) when invol attempts to reformat the drive. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)