ProBE with OS 6.4.2 locks up on Scrub

[Sandshark]

I have a ProBE, upgraded to an E7500 processor, 4GB of RAM, and OS 6.4.2.  It has 6 3TB WD red's and a LOT of files filling it about 75%. It is my media drive, but does not run any media server software and ReadyDLNA is disabled.  It does run DDClient, but nothing else that runs continuously.  It is scheduled to do a Scrub every Tuesday.  Twice now, it has completely locked up somewhere during the scrub.  The start of the scrub is logged, the end is not.  Normally, it takes about 14 hours to complete.  By "locks up", I mean no RAIDar or other presence on the network and no display if the power button is pushed.  A hard reset does gets it going again.  One scrub did successfully complete between the two that did not.

I just started a manual one, will see how that goes.  So far, so good.

Anything that I should look for as to a reason?  No other clues in the logs I can find.

[StephenB]

A tough one I think.  There are lots of possibilities (failing memory, a lock up on the SATA bus, a btrfs bug...)

A thermal issue is a long shot, but have you looked at the temps during the scrub? 

[Sandshark]

I have not looked at the temperatures, but there are no warnings about them in the log.  The manually started scrub did complete successfully.

I do wish the ReadyNASes had a watchdog timer.

[StephenB]

Maybe try some different tests and see if you can improve the odds of failure.  For instance, try copying the data volume to \dev\null ???

[Sandshark]

I shut it down before I left on a short vacation, and when it booted upon my return I saw this in the log:

Fri Aug 19 2016 14:23:53		System: ReadyNASOS background service started.
Fri Aug 19 2016 14:23:24		System: Communication with UPS 'UPS' () is OK.
Fri Aug 19 2016 14:23:19		System: Error communicating with UPS 'UPS' ().
Mon Dec 31 2001 19:01:24		Volume: Volume data usage exceeds 70% of size (13.6 TB).
Mon Dec 31 2001 19:01:21		System: ReadyNASOS service or process was restarted.

I'm assuming that the Dec 2001 date is telling me the BIOS backup battery is dead, which isn't terribly surprising given the system's age.  Are there non-default settings that I should change when I replace the battery, and might that be a contributor to this problem?  I do have the VGA header to be able to do so.

One of the fans is also has a lot of bearing noise when at top speed before the OS is in control, but there is no sign it has actually been out of control or the unit has overheated.  Still, I need to figure out if it's the CPU or system fan that I heard and get it replaced before I lose it.  Is there an easy way to get them to spin up individually so I can assess which is the problem?  In OS4, you could do a re-calibrate; but I don't see anything in OS6.

[mdgm-ntgr]

You would just need to replace the coin battery. The system should automatically get the time from a time server if it's connected to the Internet.

If you are comfortable with SSH you could do some tests using that.