During the last two or three sync checks I experienced a lot of sync errors. SYSLOG (attached) doesn't show any errors. Memory is ok. Cables and cards are attached correctly. SMARTCTL and REISERCHK look ok.

 

Currently I'm looking in a read difference during the sync check. In the screenshot attached you can see two different read dimensions (marked). The first seven disks have a lot more reads than the next eight disks. The snapshot was taken in the middle of the sync check. So the read value should be the same because the end of the smallest disk was not reached at this point.

 

Can somebody explain this to me please? I thought that all drives are read from the beginning to the end and the values should be the same.

 

It looks that the first seven disks are attached to the mainboard (6+1) and the eight disks are attached to the two adapter cards.

 

BTW, this is a Limetech server - approx. 3 years old (some disks changed).

 

Many thanks in advance.

Harald

Run smartctl reports on all the drives and do an overnight memory test.

Run smartctl reports on all the drives and do an overnight memory test.

 

Thanks for your answer. Did it already (see original post).

 

Does anybody know why the amount of read differs for identical drives during a parity check? Are the drives not read entirely?

 

Thanks

Harald

 

Does anybody know why the amount of read differs for identical drives during a parity check?

 

That's an interesting question.

 

Would somebody venture a guess?