My NAS4Free occasionally freezes - can someone help?

[fallen2109]

Since I have switched to ZFS (RaidZ1 - 3 x 3 TB) my server freezes occasionally during the night. Usually when
this happens a scrub and/or local rsyncs are running. As nobody uses the server - the LAN switch to which it is connected
is also switched off. Then in the morning when I switch on the LAN switch and try to access the server I find it either:
1) Completely shut down
2) Still running (access only via the local console), but completely unresponsive - e.g. I can not go in shell mode
or shut it down - even not via the power switch on the server front panel.

Today for example I have fount it in state 2 (LAN switch was intentionally left ON during the night, but my my
internet router was switched off - so still no internet connection was present at the time of freeze up - 00:59 according
to the log.

The last messages from the log of yesterday are shown below:

==============================================================================================================================
Feb 28 00:59:07 nasvas1 kernel: sonewconn: pcb 0xfffffe0170288000: Listen queue overflow: 8 already in queue awaiting acceptance (3 occurrences)
Feb 28 00:57:07 nasvas1 kernel: sonewconn: pcb 0xfffffe0170288000: Listen queue overflow: 8 already in queue awaiting acceptance (2 occurrences)
Feb 28 00:55:08 nasvas1 kernel: sonewconn: pcb 0xfffffe0170288000: Listen queue overflow: 8 already in queue awaiting acceptance (1 occurrences)
Feb 28 00:46:32 nasvas1 msmtp: host=smtp.ziggo.nl tls=on auth=on user=mg.vassilev@quicknet.nl. from=mg.vassilev@quicknet.nl. recipients=my.vassilev@quicknet.nl. smtpstatus=550 smtpmsg='550 Recipient e-mail address invalid / Ontvanger e-mail adres bestaat niet' errormsg='recipient address my.vassilev@quicknet.nl. not accepted by the server' exitcode=EX_DATAERR
Feb 28 00:46:27 nasvas1 msmtp: host=smtp.ziggo.nl tls=on auth=on user=mg.vassilev@quicknet.nl. from=mg.vassilev@quicknet.nl. recipients=my.vassilev@quicknet.nl. smtpstatus=550 smtpmsg='550 Recipient e-mail address invalid / Ontvanger e-mail adres bestaat niet' errormsg='recipient address my.vassilev@quicknet.nl. not accepted by the server' exitcode=EX_DATAERR
Feb 28 00:46:27 nasvas1 upssched-cmd: Communications with the UPS APC10003rdFl@localhost are established
Feb 28 00:46:27 nasvas1 upsmon[2057]: Communications with UPS APC10003rdFl@localhost established
Feb 28 00:46:22 nasvas1 upssched-cmd: Communications with the UPS APC10003rdFl@localhost are lost
Feb 28 00:46:22 nasvas1 upsmon[2057]: Communications with UPS APC10003rdFl@localhost lost
==============================================================================================================================


What I saw on the console this morning before forcefully switching off my server was that my 3ware 9690 controller have completed the AUTOVERIFY on device on port 3,4 and 5. These are loose disks that are not part of the main data ZFS pool - first 2 are 250 GB WD disks and the last one is a 1TB Samsung 840 EVO SSD. Just to be sure the reason is not the 3ware hardware - I have switched off the AUTOVERIFY on all 3ware devices (physical ports 0 thru 5). I am on version 9.3.0.2.1310 full booting from 120 GB Kingston V300 SSD with 16 GB of RAM and 32 GB SWAP (on the same SSD where the NAS4Free boot partition resides).


Anybody can help with this problem? It seems to me that the last unsuccessful event was the sending of the outgoing e-mail, but I might be wrong.

Thank you for your efforts and time in advance!

[b0ssman]

maybe this help identifying the culprit.
http://lawrencechen.net/2014/sonewconn- ... sten-queue

but you have not eliminated the 3ware hardware. just turning off the autoverify feature does not mean its not the culprit.