swap_pager: indefinite wait buffer

grapenuts

Cadet
Joined
May 11, 2012
Messages
3
I'm having an issue very similar to this:


Two days ago, I started getting alerts about offline sectors on one of my disks. The number of sectors in the alerts has been steadily increasing, so I have a replacement drive arriving today. However, twice now I've found the whole system unresponsive and needing to be reset via IPMI. When unresponsive, the "swap_pager: indefinite wait buffer" messages are spammed to the physical console. I wasn't able to log into the GUI or on the physical console to get any more information. Even trying to reboot from the menu would hang.

With the disk errors appearing right before these hangs, it seems evident that the swap space on the affected disk is likely causing OS issues and when I replace the disk it will resolve. The question I have is: is this expected, or is there something I have misconfigured? Under Advanced settings, the swap size is set to 2GB, and I don't see anything else related to system swap. It just seems out of character for a TrueNAS system to be completely crippled by one bad disk.

Any insight is appreciated!
 

Arwen

MVP
Joined
May 17, 2014
Messages
3,611
It is helpful to follow the Forum Rules, a link handily located at the top of this page. Specifically your hardware configuration, pool configuration and TrueNAS version.
 

grapenuts

Cadet
Joined
May 11, 2012
Messages
3
I'm running TrueNAS Core 13.0-U3.1
Motherboard: Supermicro X9DR3-LN4F+
CPU: 2x Xeon CPU E5-2630 v0
RAM: 106GB
Drives:
  • 1 PNY 120GB SSD (boot)
  • 2 Seagate 2TB Barracuda (zpool1)
  • 4 Seagate 4TB Barracuda (zpool1)
  • 4 Samsung 500GB 840 SSDs (zpool2)
  • Both zpools are two-disk mirrored vdevs
Controller: LSI SAS2008
NIC: Intel x520 SFP+

It was one of the 4TB Barracudas that failed.
 

Arwen

MVP
Joined
May 17, 2014
Messages
3,611
Hmm, with an even number of drives, you are right that a single drive should not be impacting operation. I don't have any suggestions, perhaps someone else will.
 
Top