Raid array corruption on Nvidia 680i chipset

Avast Free Antivirus / Premium Security
1

http://evga.com/forums/tm.asp?m=165343

see the above thread where issue was discovered. Avast is causing Raid access failures on raid arrays on Nvidia 680i chipset motherboards.

system specs: Vista 64bit | EVGA 680i Rev A1 | Intel Quad Q6700 | EVGA 7950GTX | CORSAIR Dominator 2GB DDR2 1142MHz TWIN2X2048-9136C5D | PC Power & Cooling Silencer 610 | 2X Seagate 750GB Drives in Raid 0 | Antec P182 Case | MCR320 Radiator, MCP655 Water Pump, D-Tek FuZion CPU Waterblock, Swiftech MCW30 Chipset Waterblock, & Swiftech MCRES-MICRO Reservoir

2

unfortunately the raid corruption and avast is probably an unhappy co-incidence.

Inspite of the fact the raid built into motherboards is a software based array (vs true hardware RAID built into server controllers), no software can directly cause this to happen as the Operating system doesn’t know there is raid going on and being presented false geometries to write to. When the O/S tried to write to Cylinder 12, section 5 the RAID controller receives this instruction and using its internal mapping writes the data to where it needs to go to provide the RAID functionality.

Its more likely that there is a bug within the 680 RAID driver for Windows that could not be handling the delay caused by antivirus software (as a really basic example). When the O/S tried to write the data to the drives, the A/V filter driver hijacks that write request, scans the files, and then passes the write request back to the next step in the O/S process and never actually does any of the writing (which is the only time there actually could be corruption as the read process never changes the file being read).

3

Definitely worth investigating, though I’ve seen before “Delayed Write Error” (error 51 in event viewer if I remember correctly) with no detectable hardware faults on an overclocked P4, Intel chipset board with Adaptec RAID 1 - which disappeared as soon as the CPU clock was reset to stock speed.

4

Is it the very latest build of avast! (4.7.1098)?
If not, try to update first, please.

5

That may be caused by no PCI lock, which is usually the case with Via chipsets.

That means with each FSB increase, the PCI is automatically higher than 33 mhz.

More likely to occur if the PCI bus is at 37 mhz or higher. You know those old school overclocking days.

nForce chipsets have a feature that prevents this problem when overclocking the bus.

You need to make sure that the PCI bus is locked to 33 mhz and the PCI-E bus locked to 100 mhz.

6

pci/e bus is locked. no this is not caused by OCing. results in hard drive corruption.
version: 4.7.1074 or 1098
i’ve unistalled it…

7

Does “or” mean that you are not sure what version it was, or that the problem occurred with both these versions?

8

There has been a similar problem in 4.7.1074, but it was supposed to be fixed in 1098.
Please make sure to try with 1098.

Cheers
Vlk

9

It’s back, evga 780i this time. Version 4.8.1435

Anti Virus programs tie into the OS at a low level and if the driver thinks it has told the hard drive to write something, then the anti virus steals it and scans it and then passes it off to the hard drive but doesn’t get the write confirmation back in time… write failure.

side note: some of you guys would probably think quicktime couldn’t crash a computer running Intel software raid, but with Adaptec BIOS. yes it does. ties into the hard disk controller at a low level for copy protection…

10

nForce chipsets are pile of crap and their drivers aren’t any better… lets wait for Vlk or igor again…