Hello Everyone,

 

In the last day or so, I have had several issues with my UnRAID server.  I rebooted the server to find BIOS reporting a link error on ATA4 (disk1) and re-seated the cable.  The error went away, but the red ball was produced. Short SMART test passes, long test is running.  Is it safe to re-activate this drive if long SMART passes?

 

Additionally, ATA5 (disk4) reported several errors last night.  Dynamix produced notices that SMART errors present, but another notice moments later would state that no errors were found.  A short SMART tests passes.  To make matters worse, the drive produced a number of errors in syslog and became unresponsive for a short while.  The SATA cable was replaced.

 

Fast forward to this morning, the red-ball still exists at boot and ATA5 is no longer producing read/write errors in syslog, but I see several SATA link resets for ATA5 and ATA6 (cache).  The cache drive did not mount at boot.  Within the last few months I replaced all of my SATA cables due to ATA link resets found in syslog, which temporarily resolved the problem. 

 

Does anyone have recommendations on how to move forward here?  I'm leaning towards power supply (Antec EA380) or malfunctioning SATA ports on my Asus P5Q.

syslog.zip

I'm leaning towards power supply (Antec EA380)

 

Could be quite possible, since you're supply is dual rail (one rail handles the CPU), and the other rail (17A) is handling EVERYTHING else.  Your 5 drives alone add up to 8.9A peak.  Add in the motherboard, any expansion cards, fans, etc and you're probably running quite close or exceeding its capacity.  And power supplies do lose capacity as they age.

Thank you for the quick reply.  I will look into a replacement power supply.

You're still going to have to deal with the red ball however (rebuild) after solving the other problems

Thank you for the recommendations Squid.  A Corsair TX750 which has a single 60 amp 12 volt rail has been installed in the server.  It is a pretty old power supply so it will likely be replaced soon, but will suffice for now.

 

Additionally, the array has been rebuilt and operating fine, perhaps better, for more than 24 hours.  No additional link errors have been found in the syslog.