Obtaining kernel core dumps


#1

I experience random crashes on unstable, so I wanted to get some useful info and try to learn something from it or at least send it to devs

after reading this thread I tried to get kernel dumps - without success so far.

some diagnostic info

boot environment now (N) … 12.0-CURRENT-up-20180105_040512 NR 2018-01-05

$ cat /etc/sysctl.conf | grep coredump

kern.coredump=1

$ cat /etc/rc.conf

...
dumpdev=label/swap0.eli
# tried /dev/label/swap0.eli

$ dumpon -l

/dev/null

$ geli list

Geom name: label/swap0.eli
State: ACTIVE
EncryptionAlgorithm: AES-XTS
KeyLength: 128
Crypto: hardware
Version: 7
Flags: ONETIME, W-DETACH, W-OPEN
KeysAllocated: 1
KeysTotal: 1
Providers:
1. Name: label/swap0.eli
   Mediasize: 4294963200 (4.0G)
   Sectorsize: 4096
   Mode: r1w1e0
Consumers:
1. Name: label/swap0
   Mediasize: 4294966784 (4.0G)
   Sectorsize: 512
   Stripesize: 0
   Stripeoffset: 3465564160
   Mode: r1w1e1

$ sudo ls -all /var/crash

total 7
drwx------ 2 root wheel 3 Dec 20 18:49 .
drwxr-xr-x 25 root wheel 40 Jan 5 04:05 ..
-rw-r--r-- 1 root wheel 5 Dec 20 18:49 minfree


#2

I asked the same on telegram as I have firefox causing kernel faults when switching tabs and I’d like to submit the cores. This thread slipped through the cracks evidently and some folks have been asked to look at this. Should help. Also, the bump this gives should help folks find it.


#3

I should say, there was an error in my configs (due to wrong copy-paste or my general stupidity):
wrong: dumpdevice
right: dump_device

it seems same applies to dump_dir

I’ll report back if I can access a dump via /var/crash