Topic: "APIC error on CPU0" when i copy to raid 5

[Gert]

This is a follow up on http://forums.contribs.org/index.php?topic=37096.0

Even though my file system is clean my files still become corrupt.

I use SME for a file server only. I was running a Celeron 2.5, a motherboard with a VIA chipset. 1GB DDR1 ram, gigabit nic. SME Server 7.0 installed on a 80GB IDE drive connected to hda, and 5 identical 250GB IDE drives connected to hd[cefgh] with software raid5 (hdc being secondary master on the onboard controller and the rest connected to an adaptec IDE controller)
Everything ran smooth for months then my files started becoming corrupt. Eventually I ran fsck which took about 110 hours to complete and threw everything in the lost+found, but apparently fixed the file system. But still my files became corrupt.

I’ve got 2 ibays on the system. One linked to the raid 5 array (storage) and the other to the 80gig (temp). I watched the log files while copying files through the network to the server. When I copy to temp everything is fine but when I copy to storage dmesg spits out "APIC error on CPU0: 00(60)" and "APIC error on CPU0: 60(60)". And when that happens, the CRC of the file being copied changes and the file is rendered corrupt. I have searched the internet and the forums with no luck. All I could see is that most others with this error are running duel CPUs but I only have one.

Yesterday I replaced nearly everything with brand new hardware. Now running a LGA775 motherboard with an Intel chipset, Pentium-4 3 GHz, 1GB DDR2 ram (2x 512 in dual mode). I also replaced the network card and the 300watt power supply with an expensive 600watt. I also replaced the 80gig IDE with a 80gig SATA for my OS. Installed SME 7.1 from scratch and did a complete yum update to 7.1.3. Hooked up my other drives, started the raid5, mounted the ext3 filesystem and linked it to an ibay. AND STILL HAVE THE SAME PROBLEM!

Everything was running fine for months and I had another system set up the same way which is still running fine. What can be wrong with this one? Fsck reports the filesystem to be clean, even with a forced check. Should I copy all 850gb off, reformat the filesystem and copy it back? What can cause the problem?

PLEASE HELP!!!

[Gert]

It even happenes when i copy from the 80GB to the raid 5 array directly. (Not through the network)

[Stefano]

installed on a 80GB IDE drive connected to hda, and 5 identical 250GB IDE drives connected to hd[cefgh] with software raid5 (hdc being secondary master on the onboard controller and the rest connected to an adaptec IDE controller)

just a question...

are all the hds master or are they master & slave?

AFAIK two hds on the same ide channel is not a good choice for raid..

IMHO you'd better try a (expensive) 3ware ide controller with 8 channel..

my 2€c

ciao

Stefano

[Gert]

Master and slave. I know this is not the best way of doing it, I am thinking of replacing all my drives with SATA. But still I don't think that is related to my problem.