I am trying to use FreeBSD with the 6200/6400 series ide raid cards from 3ware and have had much difficulty getting them to run properly t raid level 1 just basic mirroring. The problem that we are having is that we cannot properly install FreeBSD 4.1-Release with the 6000 series cards. If we setup a raid 1 mirror, shutdown the machine and then disconnect one of the drives and then boot up again everything is fine running from a single drive, we can setup the system no problem get everything running all fine, but as soon as we add in the second or additional drives we start to get errors. The errors start to appear about a couple minutes after initial startup and as the drives begin to sync up their data updating the new drive. After a little while the system starts to core dump everything pretty well that loads up. Systems being used are super socket 7 motherboards atx, amd k6/2 500mhz, 128-256mb ram, 15GB hard drives (fujitsu 7200rpm), generic pci vga cards. How-To-Repeat: Just setup a raid 1 mirroring with the 3ware 6200/6400 cards and use system like normal. The problem has occured identical on 4 seperate setups of cards and motherboards.
Responsible Changed From-To: freebsd-bugs->msmith Over to twe maintainer.
http://www.freebsd.org/cgi/query-pr.cgi?pr=21686 I have done some quick further testing of this problem, and the problem has gone away as far as I can tell with the latest bios that is available from 3ware. So far I have not run into any further problems, after reflashing to the latest bios everything seems fine. But when running with the old Bios on 4.1-Release problems keep poping up. -- Chris Demers (admin@govial.net) (Web Mail : cjdemers@netscape.net ) cdemers@wincom.net (Web Alias: bill_the_cat@usa.net ) =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= COMMAND.COM not found. Running NONSENSE.COM. Ufdsk (Q/L)?
State Changed From-To: open->closed This issue is an off-by-one error in the 3ware firmware which strikes when you completely fill the card with commands. 3ware have fixed it in their more recent firmware, and current versions of the driver avoid filling the card completely as well. Thanks to Chris for following up and confirming that this was the issue.