RAID problems, mandrake 10
 

Hi everybody,

Last week I installed a fresh new mandrake 10 official version on my PC. Hence my PC has only two ide buses I upgraded my Pc and added to PCI slot EIO ATA133 expansion card with two more IDE buses.

Also I bought a Enermax power supply with output power 300W.

I wanted to create two raid arrays, two mirrors: 2x80GB and 2x120B.

The system is on another 8GB disk with CD-ROM.

When I setted up the RAID arrays everything was ok, I formatted them using mkfs.ext3 and then mounted.
But there was a problem> everytime when the raid started the dma was setted to 0 on one of the two disks. (hdh 0, hdg 1, hde 0 hdf 1) And the copy speed was very slow. The CPU usage was very high. I found on the forum that it is because the bus is everytime reseted, and there is a CRC error. I changed my cables and now is everything ok (maybe the cable was too long ? (90cm) )
But I dont beleive that the cable was the problem because now I have a problem again.
In /var/log/syslog I noticed these errors:

Aug 13 15:42:37 Server kernel: md: could not lock hdg1.
Aug 13 15:42:37 Server kernel: md: could not import hdg1!
Aug 13 15:42:37 Server kernel: md: autostart unknown-block(0,8705) failed!
Aug 13 15:46:45 Server kernel: udf: registering filesystem

What does it mean? md could not lock hdg1 ? What is import ? Please if you can help me, I would be very glad. Is there a hardware (poor power supply, again bad cables, or too much devices) or software problem, ?
Will my disk gona die - will I loose my data ?


1000xx Thank you for you advise

Tomas.

Please if this question or problem was discussed send me a link, because i didnt find any solution on the web yet.