mea@sparta.com (Mike Anderson) (12/30/90)
We just got a copy of SunOS 4.1.1 here for our SparcStation 1+ and we are seeing some bizarre behavior when we try to install it. First, 4.1.1 now can perform synchronous SCSI transfers (4.16 MB/s !) to disks, and I think that is where we are getting into trouble. We are using 1.2 GB Fujitsu 2263S/H SCSI drives and CDC Wren IVs. When the Miniroot comes up, it correctly identifies these disks and says that it will talk to them in synchronous mode. I can install the miniroot on sd(0,0,1) (the Fuji swap partition) and everything seems to go just fine (I install from 150MB Cartridge tape). When it then tries to reboot the just installed miniroot, it craps out with either "Invalid instruction" or "Level 15 interrupt". The Fuji drive has a jumper that enables synchronous transfers and the jumper was installed when I got these messages. I have also tried it with SCSI 2 enabled on the drive, but to no avail. When I tried to install 4.1.1 without the Sync transfer jumper installed, the install seemed to go O.K., but I started getting all kinds of READ errors from various blocks on the disk (that is when I tried the sync jumper). I've also tried reformatting the drive in the various configurations (no sync jumper; sync jumper; sync jumper and SCSI 2; etc.) and still no luck. Has anyone seen this problem? Any clues as to how I can get around it? The faster SCSI transfer speed sure would be nice if it works. Otherwise, is there a way to install 4.1.1 but turn off the SCSI synchronous transfers in the Kernel? Thanks for any suggestions and I'll summarize to the net. Mike Anderson mea@sparta.com (703) 448-0210
rick@olivea.olivetti.com (Rick Meneely) (01/15/91)
From article <976@brchh104.bnr.ca>, by mea@sparta.com (Mike Anderson): > We just got a copy of SunOS 4.1.1 here for our SparcStation 1+ and we are > seeing some bizarre behavior when we try to install it. > When it then tries to reboot the just installed > miniroot, it craps out with either "Invalid instruction" or "Level 15 > interrupt". > [...] but I started getting all kinds of READ > errors from various blocks on the disk (that is when I tried the sync > jumper). > Has anyone seen this problem? I installed 4.1.1 on a 3/60 several days ago and all worked fine...however today (while reading this newsgroup) I was also attempting to install SunOS 4.1.1 on a Sparc 1+ and I encounted precisely these same problems!!! I am installing it on a Maxtor XT-8760S drive which is different than Mike Anderson was using - so I presume (guess) that is not specific to this make of hard disk. Any answers or suggestions would be very appreciated. | Rick Meneely | Internet: rick@olivee.ATC.Olivetti.Com|
wolfgang@uunet.uu.net (Wolfgang S. Rupprecht) (01/18/91)
rick@olivea.olivetti.com (Rick Meneely) writes: >[ I installed ] SunOS 4.1.1 on a Sparc 1+ and I encounted precisely >these same problems!!! I am installing it on a Maxtor XT-8760S drive >which is different than Mike Anderson was using - so I presume >(guess) that is not specific to this I tried installing 4.1.1 on a CDC/Imprimis/Seagate Wren VI 94191-766. Same problems. Random hangs and illegal instructions. I installed 4.1.1 by hand after untaring the tape files by hand, and building a 4.1.1 kernel from the *.o's. The resultant kernel hung just after the device probe printf's. Turning the SYNC SCSI option off in the kernel didn't help. Wolfgang Rupprecht wolfgang@wsrcc.com (or) uunet!wsrcc!wolfgang Snail Mail Address: Box 6524, Alexandria, VA 22306-0524
jamison@ccncsu.colostate.edu (Jamie Gulden) (01/22/91)
In article <1232@brchh104.bnr.ca> rick@olivea.olivetti.com (Rick Meneely) writes: >From article <976@brchh104.bnr.ca>, by mea@sparta.com (Mike Anderson): >> We just got a copy of SunOS 4.1.1 here for our SparcStation 1+ and we are > >> [...] but I started getting all kinds of READ >> errors from various blocks on the disk (that is when I tried the sync >> jumper). > >> Has anyone seen this problem? We have just spent a week with the disk error problem and (no thanks to Sun) have found our problem to be in the SCSI interface. Depending on the length of the cables, the particular terminater used, and how many devices are on the bus will affect the success or failure. We originally got our system running by removing one of the devices (leaving 2) from the bus. National Peripherals suggested shorter cables even though we were below the 6 meter limit and this worked and allowed us to put the third device back on the machine. Suns only suggestion was to remove the terminator.
bkc2@midway.uchicago.edu (Benjamin Clardy) (04/17/91)
Our main computing bottleneck seems to be disk speed. Specifically, we have SAS analyzing fairly large datasets, 5MB - 150MB, so sustained disk throughput is very important. I have tried to implement sync scsi under 4.1 changing the scsi_options in the kernel to 0x38 but I can not tell any noticeable effect in speed. In addition, if I set the jumpers on my M2263SAs to 4.0MB/s max transfer rate (instead of 2.7MB/s) the drive fails to function properly. Do I need to go to SunOS 4.1.1 to take advantage of sync scsi? What type of throughput inccrease with sync scsi can I see on a SLC? What other advantages does 4.1.1 have over 4.1? FLAME ON My internal Maxtor 207MB hard disk in my IPC just failed again for the second time in 6 months. This time the drive is not covered under Sun's whole THREE month warranty. They want $600 for a new one with a 30 day lead time, $1080 for one with a 1 day lead time. I said screw it, and went out to Fast Access and picked up the Fujitsu that I already had on order. (This is not a commercial for Fast Access, but they did help me out of a bind). The Fujitsu was $1413 for 660MB, with a SIXTY month warranty. Why is Sun selling a drive with a MTBF (at least for me) of three months? Ah, but perhaps that is why the warranty is only for three months. I am buying all of my machines diskless from now on. Long live my Fujitsus FLAME OFF Benjamin Clardy bkc2@midway.uchicago.edu