From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.2) Gecko/20010726 Netscape6/6.1 Description of problem: When a promise FastTrack66 raid controller has any of the following combinations, the pdcraid.o module or inline compiled code fails to attach the device through /devs, yet /proc/ide/PSC202xx correctly shows the configuration Invalid combinations raid 0 stripe accross 3 disks in any combination accross the 2 IDE busses raid 0+1 accross 4 disks basically any combination that uses any raid 1 Working combination raid 0 accross 2 drives located on IDE 0 and IDE 1 ports Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Load RH 7.2 (2.4.2-10) onto HDD attached to motherboard IDE 2. Confirm the devices /dev/atariad/d0* exist 3. Insert FT66 into system and attach HDD's accross the two IDE ports as stripe/mirror 4. Boot system onto motherboard HDD 5. insmod ataraid 6. nsmod pdcraid 7. *device error 8. cat /proc/ide/pdc202xx shows correct configuration Method 2 1. Compile kernel with the promise/ata code inline, not as modules 2. Reboot and watch for boot message about Promise software raid 3. Note - no raid drives found 4. fdisk /dev/ataraid/d0 == *cannot open device 5. cat /proc/ide/PDC202xx == correct configuration Actual Results: Method 1 Cannot install module pdc202xx.0 Method 2 No raid arrays detected Cannot open /dev/ataraid/d0 for fdisk Expected Results: Method 1 Module should load Method 2 Expect to see raid configuration as system boots Expect to fdisk the /dev/ataraid/d0 device Additional info:
Raid 0+1 is currently unsupported by the pdcraid driver; Raid 0 should work though. Could you please attach the output of "dmesg" so that I perhaps see what is wrong ? Also please try the 2.4.9-7 security update kernel as that has a newer version of the pdcraid driver...
Have installed 2.4.9-7 and testing as follows 4 hdds's now available under the following configs test 1 raid 0+1 test 2 raid 0 for the first drive pair ide0 master, ide1 master raid 0 for the second drive pair ide1 slave, ide1 slave results test 1 mounting of raid 0+1 array correctly performed /dev/ataraid/d0p1 test 2 mounting of first raid 0 array correctly performed /dev/ataraid/d0p1 mounting of second raid 0 array correctly performd /dev/ataraid/d1p1 error testing test 1 loss of any disk in the raid 0+1 array results in entire array being offline test 2 loss of a single disk in the raid 0 array results in the array being offline yet still accessable by mounting through /dev/hd[e-h]*
I'm suprised raid1+0 works as I didn't add any code for it. The "raid0 fails when you loose a drive" is hardly surprising....
Sorry brain slip Test 2 should read raid 1 array not raid 0 array with loss of one diskk in the mirror for error testing
we are trying to install a number of rack mounted, 1U systems. on-board Promise FastTrack100, 2 disks. chosen raid1 over both HDs. no further HDs, so we'd like to make root part on the raid1. using bootnet.img+updates-disk-20020117.img, the kernel sees 2 HDs AND 1 raid1, but anaconda don't propose the raid as an option for partitioning/installation. I suppose that I have somehow to tell anaconda to use /dev/ataraid/dXpY
I forgot to mention that of course bootnet.img comes from redhat-updates/7.2/en/os/images/i386/: -rw-rw-r-- 1 731 1474560 Jan 18 00:57 bootnet.img -rw-rw-r-- 1 731 1474560 Jan 18 00:57 pcmcia.img -rw-rw-r-- 1 731 1474560 Jan 18 00:57 update-disk-20020117.img -rw-rw-r-- 1 731 1474560 Oct 22 03:29 update-disk-20011009.img best regards
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/