several bad symptoms

[buck09]

i recently added this controller card (about 2 days ago)

https://www.superbiiz.com/detail.php?name=AOC-SASMV8

 

and added a single 2tb seagate drive to it.  after booting it was recognized and cleared and working fine.

 

last night i was about to watch a movie with my roommate and about 3 minutes into it the movie stops on my htpc and it claims it cannot connect to the server over the network.  so i stop my down/uploads, stop the array.  the array would not stop after 20 minutes, so i tried ctrl-alt-del on the server keyboard.  this did nothing, so i hit the reset button. 

 

after rebooting, which took longer than usual, the cache drive was missing.  it's an ocz ssd drive, which seems like it might be dead, but i'm about to take it to compusa to have it checked.  but i was able to bring the array online and finish watching the movie.

 

after this, i shut down the server to double check all the connections.  and reboot.  still the cache drive wasn't found.  so shutting down again i added the cache drive to the new controller card.  boot up and this time the cache drive and the new 2tb seagate are missing.  power down again, remove the cache drive from the server, plug it into my windows machine and get nothing.  leave it out of the server and boot up, and still the new 2tb drive is not seen. 

 

i love the whole idea behind unraid, but it is frustrating when everytime i reboot the server there is a problem. 

 

 

edit: running server plus 5.0 b14

syslog.zip

[RobJ]

Several comments, I'll start with one that's probably not that important - your syslog has over 6400 lines, of which about 5400 have to do with installing packages, mostly various versions of Python (and sqlite) over and over, and most of that deleting unwanted parts of the Python packages.  I'll admit I have little knowledge of these packages, but it looks very wrong, and I recommend asking someone knowledgeable with these if there are improvements available with these.  Ask them if it is possible to provide packages without all the unwanted stuff, so it does not get installed then deleted, over and over.

 

The system found 4 onboard SATA III ports (configured as AHCI) with 3 drives attached, a SiliconImage 3114-based disk controller with 4 slower SATA ports (does not support SATA II) with no drives attached, a pair of drive channels configured as IDE channels with 2 modern SATA drives attached, and belatedly an MVSAS disk controller, with no drives visible.  I suspect that the 2 SATA drives on IDE connections are configured that way by default in the BIOS settings, so you should on reboot go into those BIOS settings and try to change them from IDE Emulation to a SATA mode, preferably AHCI.  I suppose it is also possible you have them connected to IDE connectors by way of SATA-to-IDE adapters, but if so, you are losing a little performance that way.

 

It does look like it struggled to set up the SAS card, but no errors were reported.  I suggest you upgrade to the latest v5.0-rc5 version, because it has later and probably better mvsas support.  It may handle the card better.  If no better, then you probably have a defective card.

 

Other than the missing Disk 5, there are no real errors in this syslog, but that is because syslogs start fresh with each boot.  Next time, please try very hard to capture the syslog right when you are having problems, and BEFORE you try to reboot.  That is the most important syslog for troubleshooting what went wrong.  Once you reboot, it is lost.  I cannot say what is wrong with the missing drives, because the evidence is now gone.

 

The reason the first reboot took longer is that when you could not shut down correctly, and had to hit the reset button, the system begins with a parity check on the next boot, plus usually has transactions to replay.  Both processes slow the start up way down, and sometimes enough so that the setup of certain things is missed, like User Shares and the Cache drive.  If you cancel the parity check, then stop and reboot, all should be well again, but you really should run a parity check as soon as convenient.