panic: vm_thread_swapin

[pdanshov]

New FreeNAS build 2nd crash / hang after about 24 hours

Version: FreeNAS-8.3.0-RELEASE-p1-x64 (r12825)
CPU: AMD E-350D APU with Radeon(tm) HD Graphics
RAM: 16302MB Crucial Ballistix Sport Series 2x 8GB DDR3
Mobo: ASRock E350M1
Boot Device: 2Gb USB stick attached to Mobo USB Header

The first time it crashed the system was either idle or a user tried to copy some files to it over CIFS, the second time it was almost done rsyncing ~800Gb from ada1 to ada3 through WebGUI Shell.
After 2nd crash I changed RAM settings from AUTO to MANUAL and set the timings to 9-9-9-24. Also enabled CPU virtualization, disabled Combined IDE mode, and set drives to AHCI.

 

[William Grzybowski]

Sounds like you've some applications hitting swap and the disk that has the swap is having problems/starting to fail (ada0) ?

Whats if your pool configuration, how many disks? Are you running plugins jail?

 

[pdanshov]

ada0 was the one possibly being written to in the first crash, it is also the disk with the CIFS home directories.
ada0 is UFS and the other two are NTFS & UFS.

Does FreeNAS automatically setup swap for each disk?

Not sure about pool config, they are all individual disks, no raids, and no plugins jail yet though I will be installing all of them.

Also, I have extreme power saving mode set on all disks (spindown after 20 mins, APM level 1) and ada0 kept spinning up/down every 5 mins it seems like and I'm sure nobody was trying to access it, maybe SMART checks?
Or some kind of persistent windows mounted share refresh?

In any case SMART was disabled in the bios so I enabled it and the system has been running stable now for almost 24 Hours without the strange spinning up/down of ada0. Maybe due for a crash/hang soon, hopefully not.