Hi folks,
Since many days I'm trying to find out the reason for the crashes during the backup of my files. For simplicity I only speak of cp, but cpbk as well as rsync gives me problems. I hope somebody can give me a good hint since I haven't found anything helpful in the net.
WHAT I HAVE:
Hardware: Dual Celeron 450, ASUS P2BD, 384MB RAM, Promise U133TX2 (PDC20269), 40GB WD (internal IDE) + 2x80GB Maxtor + 2x120GB Maxtor HDs (Promise)
Software: SuSE 8.2 (2.4.20 SMP 64GB), Software Raid0 (/dev/hdf1 +/dev/hdh1 = /dev/md0), reiserfs (3.5 and 3.6)
PROBLEM:
"cp -av" from and to any hd connected to the promise controller gives me either a system freeze or segmentation faults (see below) (which look different each time) after a certain amount of copied data (vary each attempt from 1-2 to 8-10GB).
I also made tests without the raid, i.e. I removed /dev/md1, repartitioned and reformatted the hds and copied some files -> crash
For test purposes I connected one hd as primary slave at the internal IDE controller and started a copy of 20GB from this hd to the system disk -> no crash.
For this reason I guess it has something to to with the promise controller. However it is officially supported by the 2.4.20 kernel.
But the module seems to be loaded correctly.
MORE FACTS:
- there are no IRQ-conflicts
- tried different PCI-latency settings in the bios (32 - 0)
- also reinstalled OS (was SuSE 7.3 before)
- swapped hd cables, hds, ram
- turned on/off ACPI
- played around with the DMA settings (dma=off is awfully slow, thus I didn't test it)
- checked the hds with Maxtor's "Powermax" diagnosis tool
Unfortunately I'm not allowed to post any links to give you more information on my system
Does anybody know of any problems regarding promise and SMP?
I remember there is a possibility to temporary disable one CPU at boot-time, isn't there?
I would be grateful for any hints.
Sascha