Date: Sun, 3 Mar 2002 01:10:36 -0600 From: Bob Giesen <BobGiesen@earthlink.net> To: freebsd-questions@freebsd.org Subject: FBSD 4.4 UDMA ICRC error Message-ID: <E16hQ8o-0003kX-00@falcon.prod.itd.earthlink.net>
next in thread | raw e-mail | index | archive | help
I spent a couple of days messing around with this upgrade, so I'd really like to know what happened, here, but I won't lose any sleep over it. The box at issue is happily running FBSD 3.2 again and I don't feel compelled to upgrade it, just yet... I have a dual-boot (w98/FBSD) system on which I've run FBSD 3.2 for a couple of years, without major incident. When I recently successfully upgraded another system to 4.4, I decided to upgrade my second box. As soon as I booted it, however, I began to get many errors like this one: ad1s1f: UDMA ICRC error reading fsbn 17694848 of 8060992-8061007 (ad1s1 bn 17694848; cn 17554 tn 6 sn 38) retrying I searched the FBSD archives and Googled the Net, only to find suggestions that I had (A) a faulty hard drive, (B) a faulty IDE ribbon cable, or (C) electro-magnetic interference from nearby power cables or a coil. Hoping for a cheap and easy fix, I first tried rerouting my ribbon cable, to no avail. (I actually held it as far from everything as possible while firing up the box, but I still got the errors.) Since I have a few, never-been-used cables, I tried a couple of these -- one of each, 40- and 80-conductors. No dice. Then, I wondered if it might be the power cable supplying the hard drive, so I swapped supply lines a couple of times. Nothing doing... Fearing the worst, I still couldn't help but wonder if the upgrade to 4.4 somehow got hosed. So, I backed up my /etc and /home trees and did a *clean* install of 4.4 -- newfs'd the drive and installed from scratch. Funny thing happened when I did the backup: I tar'd and cp'd those trees to my DOS drive -- and got the same errors during the backup -- except some of them referred to the DOS slice (ad0s7) to which I was saving the tar files. Hhmmm... fat chance that both of my hard drives went south at the same time. I was now pretty sure I had an OS problem -- be it native to the new version or something in my configuration. After installing the fresh 4.4, I still got the UDMA ICRC errors -- with the original GENERIC kernel as well as with a few custom ones I compiled. I got X working and the box, for the most part, seemed to behave normally while working in the GUI, but command-line operations would repeatedly spit out those pesky error messages. Finally, frustration having set in and beginning to feel some fear of losing even more time due to some future failure, I gave in to the nagging question of how the box might behave if I reinstalled 3.2 -- so, I did. It is back to normal for about 4 days, now -- not a single error of any kind. Being that the UDMA ICRC errors began with the first boot of 4.4 and continued through the last login session on that install -- and that the errors ceased immediately upon wiping it off the hard drive and reinstalling 3.2, I have little doubt that the problem lay somewhere in the software. So, does anyone have a clue as to what might have caused this or how I might have prevented it by way of system configuration? The relevant hardware is: Maxtor 91303D6 13-GB HD's (both, ad0 (DOS) and ad1 (FBSD)) -- (33MB/sec UDMA mode 2) SCE (Superpower Computer Electronics) SP - A586B mobo w/ Award BIOS ver A.9 AMD K6-2 500 AFX I have both hard drives on the primary IDE port (FBSD slave) and my only secondary IDE device is a DVD-ROM. Grasping at straws, I tried the DVD as slave as well as master to see if it affected either 3.2 or 4.4 in any important way; it didn't. My dmesg output (which I didn't save, from any 4.4 boots) didn't seem to indicate anything wrong - except for those UDMA errors popping up as the boot prgressed. (It's probably just a coincidence, but I did notice that the error would first appear immediately after the word (linux) -- on the same line, with no preceding space -- presumably while the linux compatibility was being loaded by the kernel...) If it's of any use, I can provide a current, 3.2, dmesg output (the hardware's the same); I'm leaving it out of this message to reduce the length of this already-long message. Thanks, in advance... Bob -- "The ultimate measure of a man is not where he stands in moments of comfort and convenience, but where he stands at times of challenge and controversy." -- Martin Luther King, Jr. (1929 - 1968) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E16hQ8o-0003kX-00>