LSI RAID monitor reports "Consistency Check inconsistency logging disabled"

I have a server with a LSI MegaRAID 9261-8i controller. Recently I started getting alerts like this one:

Controller ID: 1  Consistency Check inconsistency logging disabled, too many
    inconsistencies on VD: 0
Generated on:Sat May 12 04:06:40 2012


SYSTEM DETAILS---
IP Address: 192.168.1.29
OS Name: Windows 7 x64
OS Version: 6.01
Driver Name: megasas.sys
Driver Version: 4.5.1.64

IMAGE DETAILS---
BIOS Version: 2.120.33-1197
Firmware Package Version: 12.12.0-0045
Firmware Version: 3.21.00_4.11.05.00_0x05000000

VD 0 is a RAID mirror containing the system disk.

I have searched and read, but cannot find any trace of how to actually do anything about this. I tried running a scandisk but that did not find anything (as I expected, since scandisk reads the disks as exposed by the controller, right?). The MegaRAID Storage Manager does not as far as I can see have any options for checking or fixing physical disks. The program claims the VD is "healty", and both disks have Error count 0.

Also a bit strange is the System details in the message... The IP address is associated with the RAS (dial in) interface, and the OS should be Windows Server 2011 SBS.

Has anyone else experienced this before? What can be done?


Don't sweat it. For mirrored disk or RAID 1 the consistency check message is NOT an error message. It is something that occurs thousands of times per second in the system memory subsystem and cached memory, where it is called cache coherency and is never reported. RAID1 consistency check is equivalent to memory cache coherency.

The consistency check function is a synchronization process. The function of the consistency check is to make sure data on the primary disk and its replica match. The method used to do this is to generate parity for a stripe of data in both the primary and replica, then it compares parity. This technique is effective and much faster than trying to compare each byte of data. So a message, "Consistency Check found inconsistent parity on VD 00/0 at strip 33617" is "technically" correct, but for RAID1 is NOT cause for concern because it does NOT have the same meaning as a data parity error in RAID5 or RAID6. Since you reached the programmed threshold, CC logging was stopped.

A RAID1 consistency check is a synchronization mismatch. If a consistency check is detected, it is simply stating that the replica has stale data at the moment. Inconsistent data, by itself, is in no way a hard drive failure. It simply indicates that the fresh data on the primary disk has not been put on the secondary disk yet. In the absence of Patrol Read failures, the disks are good and the data is safe. There is NO performance problem.


According this IBM document it can occur on RAID1 or RAID10

https://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=migr-5092480

The above document says:

Reduce disk paging by:

  • Using a cached IBM ServeRAID adapter
  • Increasing the amount of system memor