Mac Pro Bay 2 disk causing Video Problems?

[zoomp]

Hey Guys,

I am writing this here since I have not seen anything like this before anywhere and I still do not understand what exactly could be causing this issue.

I booted my 2010 MP today, and to my surprise both my 2 screens were flashing and after sometime it would hang on with large blocked reddish image. At first, I thought my nvidia GTX 970 was starting to die, which made me a lot worried.

After a lot of troubleshooting I found out a HD on Bay 02 is dead (I got some SMART warnings this past week) and it is causing it. It was used on a SoftRaid Raid-5 with Bays 1,3,4 (and 2).

So, I can't really see the logics behind it. Would you care to take a good guess of what could be causing this weird correlation between GFX card and a dead HD?

I updated to SoftRaid's latest driver anyway just in case it was causing it.

Cheers,

Ed.

 

[AidenShaw]

Hey Guys,

I am writing this here since I have not seen anything like this before anywhere and I still do not understand what exactly could be causing this issue.

I booted my 2010 MP today, and to my surprise both my 2 screens were flashing and after sometime it would hang on with large blocked reddish image. At first, I thought my nvidia GTX 970 was starting to die, which made me a lot worried.

After a lot of troubleshooting I found out a HD on Bay 02 is dead (I got some SMART warnings this past week) and it is causing it. It was used on a SoftRaid Raid-5 with Bays 1,3,4 (and 2).

So, I can't really see the logics behind it. Would you care to take a good guess of what could be causing this weird correlation between GFX card and a dead HD?

I updated to SoftRaid's latest driver anyway just in case it was causing it.

Cheers,

Ed.

Just a guess - but consumer hard drives have aggressive error recovery features that will retry errors for 30 seconds or so before giving before giving up and returning an error to the OS. On many systems these retry scenarios can block all I/O on the system until the drive gives up. (Windows, Linux and a long list of other systems that I've used.)

If you have a system with random hangs of 20 to 30 sec - first thing that I'd check is to see if the error log shows disk errors. Interrupting the data flow to a graphics application could result in screen artifacts.

Never ignore a notification of a SMART warning or error (although looking at the raw SMART counters you might see some scary things that are harmless).

My HPE ProLiant servers' RAID controllers are set for "predictive failure replacement". If SMART suggests that a drive is in trouble, the drive will be ejected from the RAID set and replaced with a hot spare. I'll get a message that I can send to HP support and they'll send me a replacement drive. Note that the drive has not yet failed, but SMART predicts that it will. That's good enough for HPE to send me a new one.