pid 63342 (pkg), uid 0: exited on signal 11

[aplats]

I get this message once a night, only one line, so it looks like it's only happening once. It's always pkg application that exits.

So, doing research I see this can be a memory problem. However, I have ECC ram installed and I did Memtest before even installing FreeBSD. Memtest did the bit fade test which checks to see if any bits get flipped over some time in the RAM. Everything passed fine.

I get no other errors than that one. I also don't remember getting these errors with FreeNAS 9.1, but I didn't have it installed long before using the GUI to upgrade to 9.2. Also, I don't get the messages every single night. So far I received it on Dec 25th, 26th, 27th, and then the 31st. Nothing from the 16th through the 24th, though. I switched to FreeNAS 9.2 the day it came out on the 22nd.

It would really blow if I got faulty hardware, but it seems like I should have more things crashing than just pkg once a night if that was the issue. Also, I guess memtest won't get everything, but it did pass fine on the run I did.

Edit: Another thing I forgot to mention is that this is the Kingston Server Value RAM. It's supposedly tested pretty well to ensure the RAM is working perfect.

I also have a 9.1 Jail running Samba4 for Active Directory.

Specs of my FreeNAS box:

-Seasonic Platinum 760 watt power supply (760XP2) hooked up to a 860watt "Pure Sine Wave" CyberPower UPS, since some power supplies do not like the modified square waveform.

-Xeon 1245 V3 using stock cooler, which should be plenty for this machine

-Asrock C226 WS motherboard

-16 GB Kingston Server Premier ECC unbuffered RAM

-2 x Seagate 4tb NAS drives for an encrypted mirror vdev

-1 x Seagate 4tb NAS via an eSATA enclosure (using encrypted ZFS) for rsync every 2 days

-1 x Toshiba 128GB Q Pro SSD which has the OS installed on it (I know a flash drive would work just as well, but I had this drive being unused for now, so I used it)

-All in a Fractal Designs R4 case with 2 x 140mm intake and 1 x 140mm exhaust, which should provide great cooling

 

[aplats]

What I find strange about this is it's only the pkg process that this error happens on. I just put up a new Jail (my old one was FreeBSD 9.1 running on my 9.2 freenas, new jail is now 9.2) for Samba4, I have compiled Samba4 and its dependencies twice now, and both times ran fine. I would think that if the memory or motherboard was bad, this problem would happen on something other than just the pkg process.

I went ahead and re-seated the memory, as well as move it from slots 2 and 4 to 1 and 3. I shutdown the old 9.1 jail and am now using the 9.2 jail for Samba4 Active Directory. I have about 7 days to exchange the motherboard or RAM if there is a problem, so hopefully I can figure something out before then.

 

[dlavigne]

Did you get this figured out?

 

[aplats]

Still working on it. It's just pkg and when it does happen, it happens at 3:01am. So it looks like pkg is being run with the periodic daily task?

If I try to run the pkg command on FreeNAS, it's not even initialized yet, so I can't use it. Not sure what it's being used for on the periodic daily, as looking through /etc/periodic/daily, I can't find any script that calls pkg (but I did find one that calls pkg_info).

I had pkg exit out of a Jail yesterday morning at 03:01am for signal 11 also. PKG is initialized on that jail and works just fine whenever I call it. However, I did notice for some reason it was version 1.2.3 instead of 1.2.4_1. I have no idea why that's the case, because every other jail I roll out has 1.2.4_1. I went ahead and recompiled pkg from ports on that jail--we'll see if it happens again.

Does it really seem plausible to be a hardware issue if it's just pkg that exits on signal 11 and only when it's run during whatever 03:01am task?

I need to decide if I am going to go through the hassle of taking out and exchanging my motherboard (or ram) for this. I got deals on both pieces of hardware, so I'd be out a box until I get them returned, verified and reshipped from newegg.

I put up a splunk server in a jail to start collecting shaking events from FreeNAS and my samba4 jail.

Every time I run something or compile things, they seems to in great, and I've compiled quite a few programs.

Do you have any ideas?

Thanks,
J

 

[dlavigne]

It may be worthwhile to create a bug report at bugs.freenas.org (and post the # here). A dev may have some insight on what's happening which is triggering the event.

 

[aplats]

Thank you for the suggestion. I'll do that right now. Hopefully I can get this figured and, and more so I hope it's just a bug in the software.

 

[aplats]

I went ahead and created bug report #3829: https://bugs.freenas.org/issues/3829