Unsolved
This post is more than 5 years old
11 Posts
0
15307
SCSI I/O errors after cold start
Hi
I have a Dell 1950 PE which has PERC 5i RAID controller and a number of disks on which the OS is installed. This works fine.
I'm trying to attach an EonStor A08U-G2421 RAID array to it. To connect to it, I added another U320 SCSI card (non-RAID). The drivers installed were the default Windows drivers (Windows 2008 R2). The RAID array is configured as an Enclosure (this is how we have it configured under Windows 2003 on another non-Dell server) . The OS could 'see' the volume although it was initially offline (in disk manager). I manually set to online and can access it as a drive and read and write to it successfully with no errors.
I ran Windows update a number of times during which the server was restarted. It continued to operate normally. However, as soon as I power down the server and array then start from cold (array first then server) after a short period of activity I get SCSI errors followed by a SCSI reset. From the event log entry the first code in the data is 0018000f.
I have tried a number of different SCSI cards. One being an ATTO PCIe card which had a driver for the OS but that too failed.
Can anyone suggest what I may need to do to resolve this?
Craig
DELL-Chris H
Moderator
Moderator
•
8.4K Posts
0
February 5th, 2014 07:00
Cnicho,
The issue you are running into is the U320's inability to understand raid, that's what is causing the SCSI communication issue and reset. The card you would need in the 1950 is the Perc4e/dc. It is the external raid controller. Now an issue you may have is availability, as Dell no longer sells them you would need to go 3rd party.
Let us know how it goes.
cnicho
11 Posts
0
February 5th, 2014 09:00
I'm not sure I follow. We have this arrangement in another non-Dell server. It too has it's own RAID controller and set of disks and an EonStor is connected to it using another non-RAID SCSI card. The EonStor has the volumes already defined. The connected server is not expected to see it as a RAID device and shouldn't.
Are you saying other non-RAID U320 SCSI cards will not work in the 1950; only a PERC 4 will?
DELL-Chris H
Moderator
Moderator
•
8.4K Posts
0
February 5th, 2014 11:00
Cnicho,
Both the LSI and Adaptec U320's are compatible in the 1950. I was under the assumption that this was more of a JBOD type configuration. After further research on the EonStor I see that this is the described setup. In regards to the SCSI errors and reset I read that the internal SCSI terminator can reset at the end of a Firmware update, was one performed recently?
Also, what is the specific error and event ID you're seeing? This can help us ID what is causing the SCSI issues.
cnicho
11 Posts
0
February 6th, 2014 03:00
I seem to have stabilised it now and the errors are no longer appearing. I needed to apply some firmware updates. Some of the versions were quite old.
After that the disk was offline and I discovered I needed to set the SAN policy in DISKPART to OnlineAll. I also needed to set the disk read-only setting to no in DISKPART. Both of these needed setting after the firmware updates; I'm not sure why. Also, another warning that was appearing in the event log before (i.e. Event Id 28 from VMSMP abount MAC address pinned to port ...) has also stopped.
I should have checked the firmware to start with. Anyway, thanks for taking the time to answer.