|
Top-level
 i bet it is this issue. basically the ZuluSCSI returns values for disk geometry that somehow confuses the Solaris partition tool. one solution is to use a real SCSI hard drive and then dd it over to the ZuluSCSI once the install is completed. https://github.com/ZuluSCSI/ZuluSCSI-firmware/discussions/122#discussioncomment-4418076 126 comments
 Many years ago I wrote a little tool called "scsi-ping" which gives you a working disktab entry for a SCSI disk. It's still out there somewhere. zululog.txt has the smoking gun: [996859ms] WARNING: Host used command 0x1A which is affected by drive geometry. Current settings are 63 sectors x 255 heads = 16065 but image size of 2097152 sectors is not divisible. This can cause error messages in diagnostics tools. i think the solution here is to resize the disk image file so it is divisible by 63*255*512. @tubetime This all reminds me of my days with @Migueldeicaza helping him port Linux (him the kernel, me the rest of Red Hat) to the SPARC. The hardware had some idiosyncrasies but we USUALLY found them to all have fairly good reasons. Usually. My favorite thing was getting Linux to install to one from tape. AFAIK I’m the only person to ever do that. that wasn't the solution lol :blobsweats: there's no more zululog.txt error but the superblock isn't getting written correctly. 6D B6 DB 6D is the test pattern written by the format command. could be somehow corrupting it during the write? @tubetime this FF is suspiciously looking like high-Z.  @tubetime 6D B6 DB is a worst-case test pattern for MFM encoding, so the NAND flash chips in your SD card must not have enough timing margin for the channel code. well, I broke down and put in a spinning rust disk. the installer is much happier now and is copying files.  spinning rust refuses to boot when the SCSI ID is set to 0. I can set it to 1 and it will boot partway but the fstab is configured for 0 so it won't work. I imaged the drive using the ZuluSCSI initiator mode, but even with the ZuluSCSI the firmware refuses to boot from SCSI ID0. weird.  @tubetime the one part of the spinning rust drive you didn't want it to emulate @tubetime might not be helpful, as it’s a fairly different machine, but I recently watched this video of someone getting a SPARCclassic up and running off a bluescsi v1, and they seemed to have a good reason to use scsi id 3 https://youtu.be/SMz2y-wdbzs @tubetime I have a vague memory of some early Sun stuff usurping SCSI ID 0, but can't recall the details.  so i tried placing a drive image at SCSI ID 3 and *it boots as SD(0,0,0)vmunix* wtaf lol. seems like Sun cheated and swizzled some SCSI IDs around. anyway. trying an older Solaris installer and it is really surprised at the date -- could it really be 10,405 days after 1995?  @tubetime the network is the computer, and the computer is a time machine  @tubetime yeah so external drives used to ship with all the little jumpers off (ID=0) and if you plugged that into your Sun4 it wouldn't boot and then you'd call the service department ... so they wanted to make the main disk some other ID but everyone assumes the boot volume is sd0 ... so some evil genius came up with this plan. The *good* external drives had a little pushbutton wheel thingy on the back to change their ID :-) SunOS 4.1.4 says it can't possibly be the year 2023: "WARNING: preposterous time in filesystem -- CHECK AND RESET THE DATE!" sorry SunOS, there's nothing i can do to fix 2023. @tubetime I booted a NeXTstation in Sep 2022 and just had to check whether I tweeted this preposterous message or tooted it. Didn't know SunOS does it too!  @tubetime I agree with the sun engineers. Why is something with a 20MHz CPU and 1gb of spinning rust (probs not the original, much more likely to ve 500MB) even powered on in 2023? @tubetime The best bit about this message is that I guarantee you there are still Sun4s in production somewhere.  @tubetime yours was the best account on Twitter, and now it’s the best account on Mastodon. Entertaining, and educational. Thank you for existing. @tubetime I've been putting off replacing the nvram in both my ipx machines :(  @tubetime 1994: Bah! That's preposterous! Reminds me of this comment in leapsecs.txt in libtai: "Note for parsers: Negative leap seconds will probably never happen, but the year 10000 will happen. Please don't contribute to the Y10K problem." @tubetime Does SunOS 4 still complain about a preposterous time after you've set the system clock to 2023? @tubetime Just seeing that terminal font made me think Solaris..I think it’s still being used in 11 or whatever they’re on now?   @tubetime That font brings back memories. I can still viscerally feel the slowness!  @tubetime This is probably the code triggering the warning: https://github.com/csrg/original-bsd/blob/master/sys/sparc/sparc/clock.c#L403-L408 It seems to be checking for dates too far in the past (before 1975, by the look) rather than too far in the future. I wonder if it is reading a zeroed sector on the disk where a unix timestamp would be in a running system?  @tubetime The condiments of the season and a preposterous new year! https://www.youtube.com/watch?v=Dv0hrQYbTXM&t=775s @tubetime the sort of thing that I will say out loud in the lab, “and this is why I have gray hair!” 🤣  @tubetime as far as I remember this is also on the SPARCstation 2, SLC, ELC, IPC and IPX @tubetime Oh, yeah, that rings a bell, zero and three are swapped in Sun4 ...  @tubetime Isn't the boot drive SCSI ID set to 3 on early Sun systems? |
Gregory's Server
@tubetime damn, I wonder if it’s some kind of Mac-compatibility that’s breaking on SPARC?