![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Looking for advice, or anyone who has experienced similar.
System: Tyan 4-way opteron (not a dell)
OS: w2k8 R2
Controller: Perc 5/e, with 256mb and bbu. Firmware 1.03.50-0461
Config: Data drives. 8 sata disks configured as raid 1 and raid 10.
Usage: These 8 spindles are the os and data drives for 5 virtual machines, but not the OS drives for the host OS.
Problem: Firmware error and all disks disappear.
So, a few days ago I notice that one of the VMs isn't responding. When i log onto hyper-v, I see that two of the guests are down, but the others appear to be running. When i try to start one of them, it fails with a disk error. When I open windows explorer on the host, all of the controller based disks are missing (worried). Same thing in disk manager (oh no).
The LSI MSM can't connect to or find the controller... The windows event log shows two related errors.
1. Fatal firmware error: Line 1116 in ../../raid/verdeMain.c
2. The driver has detected a device with old or out-of-date firmware. The device will not be used.
(Oh F*&%)
After checking a few things and doing a few searches, I bite the bullet and reboot...
Everything works after the reboot. All of the disks and logical volumes are there. All of the VMs are there. All of the VHDs are there, and none are corrupt or invalid.
After the reboot, the LSI MSM log does report a few more details.
1. About 1.5 hours before the fw crash, a read patrol had started.
After the reboot
2. all of the disks were detected as being hot plug inserted.
3. "Cache data recovered successfully"
4. Patrol Read resumed
5. Battery temp normal.
6. Battery charge above threshold.
Questions
What the heck is going on? Bad controller.. bad disk... bad cache ram.. Bad battery?
Why didn't I wind up with corrupted disks in the VMs? The crashed controller basically ripped running OS drives out of the VMs, but everything is ok... Did the battery backed write cache on the controller save my bacon?
P.S. I've disabled patrol reads on the controller for right now.
System: Tyan 4-way opteron (not a dell)
OS: w2k8 R2
Controller: Perc 5/e, with 256mb and bbu. Firmware 1.03.50-0461
Config: Data drives. 8 sata disks configured as raid 1 and raid 10.
Usage: These 8 spindles are the os and data drives for 5 virtual machines, but not the OS drives for the host OS.
Problem: Firmware error and all disks disappear.
So, a few days ago I notice that one of the VMs isn't responding. When i log onto hyper-v, I see that two of the guests are down, but the others appear to be running. When i try to start one of them, it fails with a disk error. When I open windows explorer on the host, all of the controller based disks are missing (worried). Same thing in disk manager (oh no).
The LSI MSM can't connect to or find the controller... The windows event log shows two related errors.
1. Fatal firmware error: Line 1116 in ../../raid/verdeMain.c
2. The driver has detected a device with old or out-of-date firmware. The device will not be used.
(Oh F*&%)
After checking a few things and doing a few searches, I bite the bullet and reboot...
Everything works after the reboot. All of the disks and logical volumes are there. All of the VMs are there. All of the VHDs are there, and none are corrupt or invalid.
After the reboot, the LSI MSM log does report a few more details.
1. About 1.5 hours before the fw crash, a read patrol had started.
After the reboot
2. all of the disks were detected as being hot plug inserted.
3. "Cache data recovered successfully"
4. Patrol Read resumed
5. Battery temp normal.
6. Battery charge above threshold.
Questions
What the heck is going on? Bad controller.. bad disk... bad cache ram.. Bad battery?
Why didn't I wind up with corrupted disks in the VMs? The crashed controller basically ripped running OS drives out of the VMs, but everything is ok... Did the battery backed write cache on the controller save my bacon?
P.S. I've disabled patrol reads on the controller for right now.