ff is killing named

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

ff is killing named

Christopher Barry
Hi all,

not sure if this is the right list, if not please advise.

I run my own nameserver on my box, and have for many years.
When running firefox (rebranded to iceweasel in Debian) v31.1.0, and
hitting certain websites/pages (most noticeably amazon or ebay, but it
can pretty much randomly be any site), named (BIND 9.9.5-4.1-Debian)
will gag and fall over with the following syslog dump:


Oct 11 16:45:19 monolith named[26038]: mem.c:1321: REQUIRE(ptr !=
 ((void *)0)) failed, back trace
Oct 11 16:45:19 monolith named[26038]: #0 0x7f3f05e5322d in ??
Oct 11 16:45:19 monolith named[26038]: #1 0x7f3f040407ba in ??
Oct 11 16:45:19 monolith named[26038]: #2 0x7f3f0405102c in ??
Oct 11 16:45:19 monolith named[26038]: #3 0x7f3f05702694 in ??
Oct 11 16:45:19 monolith named[26038]: #4 0x7f3f056ae16a in ??
Oct 11 16:45:19 monolith named[26038]: #5 0x7f3f05695c29 in ??
Oct 11 16:45:19 monolith named[26038]: #6 0x7f3f05695d99 in ??
Oct 11 16:45:19 monolith named[26038]: #7 0x7f3f057097fd in ??
Oct 11 16:45:19 monolith named[26038]: #8 0x7f3f0571526b in ??
Oct 11 16:45:19 monolith named[26038]: #9 0x7f3f04061eeb in ??
Oct 11 16:45:19 monolith named[26038]: #10 0x7f3f03a130a4 in ??
Oct 11 16:45:19 monolith named[26038]: #11 0x7f3f033e3c2d in ??
Oct 11 16:45:19 monolith named[26038]: exiting (due to assertion
 failure)
Oct 11 16:45:19 monolith systemd[1]: bind9.service: main
 process exited, code=killed, status=6/ABRT
Oct 11 16:45:19 monolith rndc[26558]: rndc: connect failed:
 127.0.0.1#953: connection refused
Oct 11 16:45:19 monolith systemd[1]: bind9.service: control process
 exited, code=exited status=1

I spent a good deal of time researching this issue, thinking it was a
named bug, and could not find anything about it. I'm running Debian
Jessie, which recently changed to use systemd (and I'm not sure it's
not related somehow).

Recently, I was trying to get to a page on ebay, and every time I
clicked this link:
http://contact.ebay.com/ws/eBayISAPI.dll?ShowCoreAskSellerQuestion&redirect=0&requested=22newcentury&iid=321536360296
named would crash hard and FF would fail to find the site.
It's 100% repeatable.

I decided to try chromium, to see if the error could possibly be related
to FF somehow, and sure enough chromium never crashes named and goes to
the link. I cannot get chromium to crash named.

I have never had this kind of issue before with Iceweasel/Firefox, and
it's very confusing what is causing it. I *really* do not want to use
chromium, so please let me know how we can proceed to isolate and
remedy this issue. I'm up for any testing you can think of.

Thanks!

--
Regards,
Christopher Barry

_______________________________________________
dev-tech-network mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-network
Reply | Threaded
Open this post in threaded view
|

Re: ff is killing named

Patrick McManus
Hi Christopher - I think you need to start with the bind devs for this one,
given that it is bind that is crashing. Maybe it has a no daemon mode (-D?)
and you can start it under gdb and get a stack trace - that would probably
help them. If they indicate that there is something we should be aware of
on the firefox side, please circle back.

Thanks!
-P


On Sat, Oct 11, 2014 at 5:32 PM, Christopher Barry <
[hidden email]> wrote:

> Hi all,
>
> not sure if this is the right list, if not please advise.
>
> I run my own nameserver on my box, and have for many years.
> When running firefox (rebranded to iceweasel in Debian) v31.1.0, and
> hitting certain websites/pages (most noticeably amazon or ebay, but it
> can pretty much randomly be any site), named (BIND 9.9.5-4.1-Debian)
> will gag and fall over with the following syslog dump:
>
>
> Oct 11 16:45:19 monolith named[26038]: mem.c:1321: REQUIRE(ptr !=
>  ((void *)0)) failed, back trace
> Oct 11 16:45:19 monolith named[26038]: #0 0x7f3f05e5322d in ??
> Oct 11 16:45:19 monolith named[26038]: #1 0x7f3f040407ba in ??
> Oct 11 16:45:19 monolith named[26038]: #2 0x7f3f0405102c in ??
> Oct 11 16:45:19 monolith named[26038]: #3 0x7f3f05702694 in ??
> Oct 11 16:45:19 monolith named[26038]: #4 0x7f3f056ae16a in ??
> Oct 11 16:45:19 monolith named[26038]: #5 0x7f3f05695c29 in ??
> Oct 11 16:45:19 monolith named[26038]: #6 0x7f3f05695d99 in ??
> Oct 11 16:45:19 monolith named[26038]: #7 0x7f3f057097fd in ??
> Oct 11 16:45:19 monolith named[26038]: #8 0x7f3f0571526b in ??
> Oct 11 16:45:19 monolith named[26038]: #9 0x7f3f04061eeb in ??
> Oct 11 16:45:19 monolith named[26038]: #10 0x7f3f03a130a4 in ??
> Oct 11 16:45:19 monolith named[26038]: #11 0x7f3f033e3c2d in ??
> Oct 11 16:45:19 monolith named[26038]: exiting (due to assertion
>  failure)
> Oct 11 16:45:19 monolith systemd[1]: bind9.service: main
>  process exited, code=killed, status=6/ABRT
> Oct 11 16:45:19 monolith rndc[26558]: rndc: connect failed:
>  127.0.0.1#953: connection refused
> Oct 11 16:45:19 monolith systemd[1]: bind9.service: control process
>  exited, code=exited status=1
>
> I spent a good deal of time researching this issue, thinking it was a
> named bug, and could not find anything about it. I'm running Debian
> Jessie, which recently changed to use systemd (and I'm not sure it's
> not related somehow).
>
> Recently, I was trying to get to a page on ebay, and every time I
> clicked this link:
>
> http://contact.ebay.com/ws/eBayISAPI.dll?ShowCoreAskSellerQuestion&redirect=0&requested=22newcentury&iid=321536360296
> named would crash hard and FF would fail to find the site.
> It's 100% repeatable.
>
> I decided to try chromium, to see if the error could possibly be related
> to FF somehow, and sure enough chromium never crashes named and goes to
> the link. I cannot get chromium to crash named.
>
> I have never had this kind of issue before with Iceweasel/Firefox, and
> it's very confusing what is causing it. I *really* do not want to use
> chromium, so please let me know how we can proceed to isolate and
> remedy this issue. I'm up for any testing you can think of.
>
> Thanks!
>
> --
> Regards,
> Christopher Barry
>
> _______________________________________________
> dev-tech-network mailing list
> [hidden email]
> https://lists.mozilla.org/listinfo/dev-tech-network
>
>
_______________________________________________
dev-tech-network mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-network
Reply | Threaded
Open this post in threaded view
|

Re: ff is killing named

Christopher Barry
Hi Patrick,

Thanks for replying.

I had just updated my box the other day, thinking there may be an
update to bind, but bind was not in the package list. Then today I sent
the message below to the list after noticing chromium had no problems.

After sending the email, I though I'd try to update again, just to see.
Sure enough, bind was an update. I updated it, and so far no crashing
going to the site that was crashing 100% of the time.

Sorry for the noise, weird issue, but it appears to be resolved now.
lesson: update and test *immediately* before sending to a list :)

--
Cheers,
-C

On Sat, 11 Oct 2014 18:48:15 -0400
Patrick McManus <[hidden email]> wrote:

>Hi Christopher - I think you need to start with the bind devs for this
>one, given that it is bind that is crashing. Maybe it has a no daemon
>mode (-D?) and you can start it under gdb and get a stack trace - that
>would probably help them. If they indicate that there is something we
>should be aware of on the firefox side, please circle back.
>
>Thanks!
>-P
>
>
>On Sat, Oct 11, 2014 at 5:32 PM, Christopher Barry <
>[hidden email]> wrote:
>
>> Hi all,
>>
>> not sure if this is the right list, if not please advise.
>>
>> I run my own nameserver on my box, and have for many years.
>> When running firefox (rebranded to iceweasel in Debian) v31.1.0, and
>> hitting certain websites/pages (most noticeably amazon or ebay, but
>> it can pretty much randomly be any site), named (BIND
>> 9.9.5-4.1-Debian) will gag and fall over with the following syslog
>> dump:
>>
>>
>> Oct 11 16:45:19 monolith named[26038]: mem.c:1321: REQUIRE(ptr !=
>>  ((void *)0)) failed, back trace
>> Oct 11 16:45:19 monolith named[26038]: #0 0x7f3f05e5322d in ??
>> Oct 11 16:45:19 monolith named[26038]: #1 0x7f3f040407ba in ??
>> Oct 11 16:45:19 monolith named[26038]: #2 0x7f3f0405102c in ??
>> Oct 11 16:45:19 monolith named[26038]: #3 0x7f3f05702694 in ??
>> Oct 11 16:45:19 monolith named[26038]: #4 0x7f3f056ae16a in ??
>> Oct 11 16:45:19 monolith named[26038]: #5 0x7f3f05695c29 in ??
>> Oct 11 16:45:19 monolith named[26038]: #6 0x7f3f05695d99 in ??
>> Oct 11 16:45:19 monolith named[26038]: #7 0x7f3f057097fd in ??
>> Oct 11 16:45:19 monolith named[26038]: #8 0x7f3f0571526b in ??
>> Oct 11 16:45:19 monolith named[26038]: #9 0x7f3f04061eeb in ??
>> Oct 11 16:45:19 monolith named[26038]: #10 0x7f3f03a130a4 in ??
>> Oct 11 16:45:19 monolith named[26038]: #11 0x7f3f033e3c2d in ??
>> Oct 11 16:45:19 monolith named[26038]: exiting (due to assertion
>>  failure)
>> Oct 11 16:45:19 monolith systemd[1]: bind9.service: main
>>  process exited, code=killed, status=6/ABRT
>> Oct 11 16:45:19 monolith rndc[26558]: rndc: connect failed:
>>  127.0.0.1#953: connection refused
>> Oct 11 16:45:19 monolith systemd[1]: bind9.service: control process
>>  exited, code=exited status=1
>>
>> I spent a good deal of time researching this issue, thinking it was a
>> named bug, and could not find anything about it. I'm running Debian
>> Jessie, which recently changed to use systemd (and I'm not sure it's
>> not related somehow).
>>
>> Recently, I was trying to get to a page on ebay, and every time I
>> clicked this link:
>>
>> http://contact.ebay.com/ws/eBayISAPI.dll?ShowCoreAskSellerQuestion&redirect=0&requested=22newcentury&iid=321536360296
>> named would crash hard and FF would fail to find the site.
>> It's 100% repeatable.
>>
>> I decided to try chromium, to see if the error could possibly be
>> related to FF somehow, and sure enough chromium never crashes named
>> and goes to the link. I cannot get chromium to crash named.
>>
>> I have never had this kind of issue before with Iceweasel/Firefox,
>> and it's very confusing what is causing it. I *really* do not want
>> to use chromium, so please let me know how we can proceed to isolate
>> and remedy this issue. I'm up for any testing you can think of.
>>
>> Thanks!
>>
>> --
>> Regards,
>> Christopher Barry
>>
_______________________________________________
dev-tech-network mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-network