Unexpected shutdowns

[Garry Ricketson]

 Running Minix 3 as a web server,
MINIX 3.3.0. (588a35b)
Aug 24 04:49:32 %nameserver kernel: Copyright 2014, Vrije Universiteit, Amsterdam, The Netherlands
Aug 24 04:49:32 %nameserver kernel: MINIX is open source software, see http://www.minix3.org
Aug 24 04:49:32 %nameserver kernel: Started VFS: 9 worker thread(s)
Aug 24 04:49:32 %nameserver kernel: e1000#0: Intel PRO/1000 MT 82545EM (8086/100f/00) at 2.0.0
====================
 It has been shutting down unexpectedly, after 4 or 5 days,  I am not sure which logs might have something that gives me
a clue as to why. 
I have looked at, /var/log/messages and "syslog",  but don't see anything that seems relevent, except:
From "messages"

- error: setsockopt TCP_NODELAY: Function not implemented
Sep  5 16:39:08 %nameserver 1 2017-09-05T16:39:08.450000+00:00 %nameserver sshd 255 - - error: setsockopt IP_TOS 8: Function not implemented:

I am not sure if this is significant of anything or not.

Any suggestions as to which logs to check ?

Another thing, there is a new file in my /home dir,  "core(number)", but How can I see /view this "core file", ?
Thanks

[r0ller]

Hi,

I cannot really help with the logs but the core file you could check either with gdb (not currently present on minix), lldb (not sure if it's available already) or addr2line (that should be at hand) but only if the program that dumped was compiled with -g option (debugging symbols preserved) will you get anything meaningful if I'm not mistaken.

Best regards,

r0ller

[Garry Ricketson]

Thanks ,
 I can try downloading it, "gdb" is available for my PC (OpenBsd),

[r0ller]

Hi Garry,

I'm afraid that won't help. You'd need a gdb compiled for minix3 which we used to have but not any more since 3.2.1.

Best regards,

r0ller

[Garry Ricketson]

Ok, thanks,
 Yes, I noticed , and was just now going to post that it didn't work.
 The tech at the server provider/hosting  has commented that it "lost connectivity", apparently it is
"powered on", and running, when I contact them and ask to re-start it manually.  I only have ssh access, it is
remote for me,..
 Last night I was looking at :/usr/pkg/etc/rc.d/apache

Enter code here...# You will need to set some variables in /etc/rc.conf to start Apache:
#
# apache=YES
# apache_start="start"

 And I saw this, not sure , but maybe it was / is related. In any event I added the variables mentioned to the rc.conf.
It seems to me that Minix 3 would be reliable as a server, of course this would depend on if the server itself is configured properly,
The problem for me, besides not knowing knowing much myself, I can not find much documentation at all on how it should be configured, seems like not
very many people have done this with Minix 3,..
  I also noticed in some search results, somewhere, and I am kicking myself because I did not take note as to where, but in any event it said something about if "reboot" is disabled, in the rc.d file, it would not re boot when shut down, something like that.  So any way, I did not see any thing about enabling or disabling reboot in the rc.d,  just the comments about "apache=YES ",...
  Any way, when I find I can not connect , and I contact the people at the server, they manually re boot. and connectivity is restored. I would like to get it
to where I do not have to contact them every few days asking them to restart it.
 One last comment, when I first started it, a couple of times I did have to use the 'reboot' command and it rebooted just fine, but the 3rd time, oddly , it said "command not found", and did not allow me to reboot,... instead of re booting, I just re-started "apache", everything seemed fine, for about a week,
then it "lost connectivity" or shutdown,  they manually restarted it.  Every thing seemed fine, but about the same time period 4-5 days later, lost connectivity again, or "shut down",....  I have not tried re-booting after I changed the rc.conf file, I suppose I should or need to for it to take affect.
 Any way, any other ideas, advise are very welcome.
thanks

On Tuesday, September 5, 2017 at 12:09:53 PM UTC-5, Garry Ricketson wrote:

[r0ller]

Hi Garry,

Though I'm not an apache (and not even a minix3) expert just a user but googling for "apache shutdown unexpectedly" gives a number of results so you may start having a look at them first ruling out the most frequent ones (e.g. blocked ports as far as I can see) one by one. Other than that as a wild guess, you may ask rockyandtheparrots.org how s/he manages his/her minix3 webserver as it seems to be running apache as well;)

Best regards,

r0ller

[Garry Ricketson]

Thanks,
 Hah, the link you posted "rockyandtheparrots" is me, that is the server.,... But you are right about doing the searches, my hopes are to find
the same one I saw before, thinking now, it might be in my "browser history",  any way thanks,

[r0ller]

Hi Garry,

That was funny about the link:D Another thing: have you checked the apache log, how it's set up/configured and the current log level? That only helps of course if we suspect that apache dumped. By the way, issuing the 'file' command for the core file (like 'file core') should tell which program generated it, have you checked that?

Best regards,

r0ller

[Garry Ricketson]

Thanks ,
 Well, 'file core' shows this, :

$ file core.20724
core.20724: ELF 32-bit LSB core file Intel 80386, version 1 (FreeBSD)Enter code here...

maybe a driver or firmware issue ?

[r0ller]

Hi,

Well, that's strange. When issuing "file hi.core" for the program I'm just developing (executable called 'hi') that generated a dump, I get (on NetBSD at least):

hi.core: ELF 64-bit LSB core file x86-64, version 1 (SYSV), NetBSD-style, from 'hi' (signal 11)

I'll try to look up a core dump on my minix3 system and check what it tells me there.

Best regards,

r0ller

[r0ller]

Well, could not find any previous core dumps on my minix3 system so checked it by writing a small cpp program calling a method on an string nullpointer:D

Indeed, issuing file for the generated core dump gives me the following without any reference to the executable that dumped:

core.482: ELF 32-bit LSB core file Intel 80386, version 1 (FreeBSD)

That's a pity:(

Regards,

r0ller

[r0ller]

Hi Garry,

It kept me curious, so I just gave a chance to hexedit (installed via pkgin) on my minix system and it seems the dump contains the file name that dumped 'a.out' in my case -see screenshot.

Best regards,

r0ller

[Garry Ricketson]

Thanks for the interest,
That is interesting, I can try installing "hexedit",  and see if it gives any more information.
Thanks

[Garry Ricketson]

Well , maybe I found some thing,
  It appears to me it was this "resize_image", which I installed , I am not positive it had anything to do with the server losing connectivity, but that is
the program the core belongs to.
 The "hexedit" works good, and it was available for OpenBsd as well, it worked fine on my PC so I was able to view it with out going online.
 I am not going to use the program , now and see if the server keeps running as it should.
I suppose I could try to duplicate it, resize some images, and see if it happens again, but would rather not , wish I had another computer here at home,
that I could try it on.
 Thanks.

[Garry Ricketson]

EnteMINI
MINIX-CORE......8.......
  .P..resize_image..r code here...