wyatt@cfa.UUCP (09/25/87)
I've been bitten by an apparent incompatibility between Ultrix 2.0 and the Emulex QD21 (with a Maxtor 4380) drive subsystem. I just got through installing 2.0 on one of our uVax II/GPX's, and it doesn't see the drive at boot time (it does see the controller). I's like to find out what others know about this. To convince myself we hadn't damaged something, I (painfully) went back and restored the 1.2 system on /dev/ra0a, and it booted beautifully. SOOOO, I remember a discussion several weeks back by Doug Wilson (dcw@ecsvax.UUCP) and Chris Torek (chris@mimsy.UUCP) about the 2.0 ra driver expecting a different `media ID' from the 1.2 one. This must be my problem, but the investigation I did at the time (calling Emulex, sending them copies of the USENET articles) elicited the response `it'll be OK'. Apparently, it's not! I vaguely remember from the times I formatted the Maxtors, that the QD21 does respond as if the drive is an RA81, i.e. it sends back `DU RA81' or something, to the `get status' inquiry. Maybe this isn't what it's really responding to the probe routine? I'm writing this late at night; I'm calling Emulex in the morning, first thing, but I wanted to see if anyone on the net has ideas. We have 2.0 source (although not yet dumped to disk), so modifications are possible. HELP! Bill UUCP: {seismo|ihnp4}!harvard!cfa!wyatt Wyatt ARPA: wyatt@cfa.harvard.edu (or) wyatt%cfa@harvard.harvard.edu BITNET: wyatt@cfa2 SPAN: 17410::wyatt (this will change, sometime) -- Bill UUCP: {seismo|ihnp4}!harvard!cfa!wyatt Wyatt ARPA: wyatt@cfa.harvard.edu (or) wyatt%cfa@harvard.harvard.edu BITNET: wyatt@cfa2 SPAN: 17410::wyatt (this will change, sometime)
sjb@piglet.UUCP (Seth J. Bradley) (09/27/87)
In article <689@cfa.cfa.harvard.EDU>, wyatt@cfa.harvard.EDU (Bill Wyatt) writes: > I've been bitten by an apparent incompatibility between Ultrix 2.0 and > the Emulex QD21 (with a Maxtor 4380) drive subsystem. I just got through > installing 2.0 on one of our uVax II/GPX's, and it doesn't see the drive > at boot time (it does see the controller). I's like to find out what > others know about this. I went through a similar hassle when upgrading from V1.2 to 2.0 on a system running on a U.S. Design VIP Subsystem using two 4380's. ULTRIX 2.0 would not install on the U.S. Design. The install software seems to depend on there being a genuine DEC controller out there(When, oh when, will DEC realize that a sizeable portion of their customers use non-DEC hardware?). My workaround was to re-install an RD53 as disk0 with the U.S. Design as units 1 and 2. I then installed V2.0 on the RD53. I then ran into another problem when trying to execute newfs on the U.S. Design. All the inodes above 512 were bad. A few hours with "fsck -y" fixed that, and I was then able to tar the system from the RD53 to the U.S. Design, re-configure the system, and boot off the U.S. Design. Tedious, but it worked. I realize I have a different controller than you, but this approach may still work. BTW, I hope you have better luck with Emulex than I did with U.S Design. They still haven't answered my complaint, and that was several weeks ago. -- Replies To: uunet!lll-winken!csustan!piglet!sjb Disclaimer: The above postings reflect only my opinions, and not those of any sanerclanguaget