Up to the point where I finally discovered you were right about this stupid error, I have been trying anything. Turning off (U)DMA and all performance improving options, through both hdparm and kernel options, just to see if I could get rid of this error. So far I've not seen much improvement. I'm still seeing lost interrupt error after lost interrupt error, though I'm not seeing udma-errors at the moment... The change of slave to master for the particular disc hasn't made a change on this issue either...
I haven't got a clue what I should be looking for now... Erm, help please!? :-(
The box in question is an old Pentium, running various flavours of 2.4 kernel in combination with RedHat 7.0... Right now it seems that whenever I wish to access (i.e. create a filesystem on) one of the discs connected to the Promise Fasttrack TX2 controller, it will end up in a lost interrupt error...
Regards,
Arjan
mathieu wrote:
The exact error I am getting now, when trying to mkfs a raid1 md device, is:
ide_dmaproc: chipset supported ide_dma_lostirq func only: 13
hdf: lost interrupt
ide_dmaproc: chipset supported ide_dma_lostirq func only: 13
hdg: lost interrupt
ide_dmaproc: chipset supported ide_dma_lostirq func only: 13
hdf: lost interrupt
ide_dmaproc: chipset supported ide_dma_lostirq func only: 13
hdg: lost interrupt
Any clue what I might be doing wrong?
hdf and hdg? How many harddisks do you use? If there are only two I suppose you put them on a seperate controller, both being master, so it should be hde and hdg.
Btw, changing dma settings with hdparm helped a few others (but also me) getting rid of a lot (dma) errors. (Those errors were different I know, anyway you might wanna give it a try.)
mathieu
_______________________________________________ Ataraid-list@xxxxxxxxxx https://listman.redhat.com/mailman/listinfo/ataraid-list
