View Issue Details

Jump to NotesJump to History
IDProjectCategoryView StatusDate SubmittedLast Update
0001475unrealircdpublic2004-01-06 00:572004-02-03 01:58
Reportercrimson Assigned To 
PrioritynormalSeveritycrashReproducibilityrandom
Status closedResolutionopen 
Product Version3.2-beta19 
Summary0001475: Seemingly random ICRd crashes
DescriptionI haven't located the cause of these crashes because I am usually connected to the IRCd when it does this so I'm booted off.
Steps To Reproduceunknown
Additional InformationJanuary 5th - backtrace
#0 0x0806d679 in find_cache_number (rptr=0x0, numb=0x8200598 "E8ðZN\t")
    at res.c:1324
#1 0x0806c7b4 in gethost_byaddr (addr=0x8200598 "E8ðZN\t", lp=0xbffff6e0)
    at res.c:420
#2 0x08070dbd in start_of_normal_client_handshake (acptr=0x8200258)
    at s_bsd.c:1349
#3 0x08070d51 in add_connection (cptr=0x818ef18, fd=14) at s_bsd.c:1333
#4 0x08071709 in read_message (delay=2, listp=0x8151d40) at s_bsd.c:1862
#5 0x08067ddf in main (argc=0, argv=0xbffff994) at ircd.c:1390
#6 0x42017589 in __libc_start_main () from /lib/i686/libc.so.6

January 4th backtrace:
(gdb) bt
#0 0x0806d679 in find_cache_number (rptr=0x82374c0,
    numb=0x8237594 "E8ðZ\214cº\004Õ5\020\205ØA7lBü\002R") at res.c:1324
#1 0x0806d7c0 in make_cache (rptr=0x82374c0) at res.c:1392
#2 0x0806d0f2 in get_res (lp=0x8120230 "") at res.c:964
#3 0x080720f8 in do_dns_async () at s_bsd.c:2559
#4 0x080714a4 in read_message (delay=1, listp=0x814e680) at s_bsd.c:1751
#5 0x08067d37 in main (argc=0, argv=0xbfffec14) at ircd.c:1371
#6 0x42017589 in __libc_start_main () from /lib/i686/libc.so.6

December 28th backtrace
#0 0x0806d679 in find_cache_number (rptr=0x0, numb=0x81d3af0 "EBgßÆ\f")
    at res.c:1324
#1 0x0806c7b4 in gethost_byaddr (addr=0x81d3af0 "EBgßÆ\f", lp=0xbfffdbe0)
    at res.c:420
#2 0x08070dbd in start_of_normal_client_handshake (acptr=0x81d37b0)
    at s_bsd.c:1349
#3 0x08070d51 in add_connection (cptr=0x818ef18, fd=25) at s_bsd.c:1333
#4 0x08071709 in read_message (delay=1, listp=0x814e680) at s_bsd.c:1862
#5 0x08067d37 in main (argc=0, argv=0xbfffde94) at ircd.c:1371
#6 0x42017589 in __libc_start_main () from /lib/i686/libc.so.6

If you need more, I can backtrace the rest of these core dumps... thanks in advance.
3rd party modules

Activities

codemastr

2004-01-06 01:16

reporter   ~0004540

[20:15] <Crimson> (gdb) p cp
[20:15] <Crimson> $1 = (aCache *) 0x81bb4b8
[20:15] <Crimson> (gdb) p cp->he
[20:15] <Crimson> $2 = {h_name = 0x2e6c6f70 <Address 0x2e6c6f70 out of bounds>,
[20:15] <Crimson> h_aliases = 0x29, h_addrtype = 943091817, h_length = 758133037,
[20:15] <Crimson> h_addr_list = 0x2d303231}
[20:16] <Crimson> (gdb) p i
[20:16] <Crimson> $3 = 0
[20:17] <Crimson> (gdb) p (&(cp->he))->h_addr_list[0]
[20:17] <Crimson> Cannot access memory at address 0x2d303231

syzop

2004-01-06 13:43

administrator   ~0004554

Last edited: 2004-01-06 13:43

 *question flood* ;)
1. Did you run a previous unreal version before? If so, which version and did it also have this problem?
2. Did you compile with any of: ipv6, ssl, or any other non-default options in ./Config or config.h (to help us reproduce this)
3. Did you do any source file modifications?
4. Which modules do you have loaded (output of /module)
5. How often does it happen? every day? every week?
6. If you have a network with other b19's... do they also have this problem or just this server?
7. do you run anything else on this server that has crash issues or mysteriously dissapearing (eg: apache, ftp, mysql)?
8. could you mail (or upload somwhere) the latest core file AND the src/ircd binary (so 2 files) to [email protected] ? [codemastr: if you want it too let me know ;p]

It looks like memory corruption, and these bugs are the hardest to trace.. so all (additional) information is welcome.

Thanks :)

[*edit typo* :p]

edited on: 01-06-04 13:43

codemastr

2004-01-06 16:23

reporter   ~0004556

Couple of those questions he already answered on IRC:
No IPv6
Other servers are fine (I believe he said he had one on Win2k and another on FreeBSD that ran fine).

Also about the other programs crashing, that's kind of what I've been thinking. To me it seems either a.) there is a specific host that however the NS is setup it is screwing up the resolver, or b.) The system's resolver is simply giving bad information.

crimson

2004-01-06 17:57

reporter   ~0004557

Last edited: 2004-01-06 17:58

 1. Did you run a previous unreal version before? If so, which version and did it also have this problem?

Yes... I believe I started around beta14 and have worked upwards. If you want me to try a fresh install that's no problem.

2. Did you compile with any of: ipv6, ssl, or any other non-default options in ./Config or config.h (to help us reproduce this)

ipv6 - no
SSL - yes, and at least 3-4 users including me use it
Ziplinks is on (I think the default is off)

3. Did you do any source file modifications?

No.

4. Which modules do you have loaded (output of /module)

[10:48:17] -home.n00bstories.com- *** commands - $Id: l_commands.c,v 1.1.4.24 2003/11/24 19:42:31 syzop Exp $ (Wrapper library for m_ commands)
[10:48:17] -home.n00bstories.com- Hooks: 13 15 16
[10:48:17] -home.n00bstories.com- Override:

5. How often does it happen? every day? every week?

8 times since December 13th. Most recent yesterday which prompted my visit to the support channel.

6. If you have a network with other b19's... do they also have this problem or just this server?

The hub and two servers are FreeBSD (unknown version), no problems
Two servers are Win2000, no problems (One of those is around 4 weeks uptime, maybe more)
It just seems to be a RedHat one, which never had a problem until beta19, and it didn't happen right away. I installed beta19 within a week of its release.

7. do you run anything else on this server that has crash issues or mysteriously dissapearing (eg: apache, ftp, mysql)?

I do run several things on this server... but I haven't had anything else mysteriously disappear to my recollection.

8. could you mail (or upload somwhere) the latest core file AND the src/ircd binary (so 2 files) to [email protected] ? [codemastr: if you want it too let me know ;p]

http://www.n00bstories.com/hosted_files/files.tar.gz

---

And on an unrelated note, I'm a "she", not a "he". :D

Thanks again.

[Edit: Edited a sentence for clarity.]

edited on: 01-06-04 17:58

syzop

2004-01-06 19:18

administrator   ~0004558

Last edited: 2004-01-06 19:28

 Thanks for all the info.
I get a 404 however when I try that url :(

edited on: 01-06-04 19:28

crimson

2004-01-06 19:22

reporter   ~0004559

That's because I forgot a subdirectory. :(

http://www.n00bstories.com/hosted_files/Unreal/files.tar.gz

Sorry about that. I checked it this time.

syzop

2004-01-06 19:31

administrator   ~0004560

Last edited: 2004-01-06 19:32

 Hmm, the core.9925 file is invalid/incomplete, it's only 10kb... should be more like 1-2Mb. Also since it's exactly '10240 bytes', perhaps you ran out of quota? if so.. then just give me the previous core dump (assuming you didn't recompile it between coredumps).

edited on: 01-06-04 19:32

crimson

2004-01-06 19:39

reporter   ~0004561

OK, I won't even pretend to know what happened there... I re-packaged it in the same location and it *should* be correct now.

crimson

2004-01-06 22:18

reporter   ~0004568

OK, we have mpatrol installed on the box per our PM conversation today.

syzop

2004-01-30 22:15

administrator   ~0004802

got any results / crashes? :)

crimson

2004-01-31 09:37

reporter   ~0004808

Ever since we installed the mpatrol it hasn't crashed once!! It's gotta be coincidence, right?

codemastr

2004-01-31 17:38

reporter   ~0004813

It's called a Heisenbug, named after Werner Heisenberg. Basically it is a bug that follows the Uncertainty Principle. The more you do to try and understand the characteristics of the bug, the more you alter the bug, and therefore the harder it becoems to find. :)

syzop

2004-01-31 17:47

administrator   ~0004814

that almost sounds like quantum mechanics [eg: if you look at something, it changes] ;)). *dinner*

codemastr

2004-01-31 18:20

reporter   ~0004815

Exactly, the Uncertainty Principle is the basis of quantum mechanics ;)

crimson

2004-01-31 22:17

reporter   ~0004825

LOL - you guys are great. I guess when I said the crashes were "seemingly random" I was right. :P

syzop

2004-02-03 01:58

administrator   ~0004853

Ok, I'll close the bug then :).
If you happen, for some reason, suddenly to be able to reproduce this bug, mail me at [email protected] with the backtrace and last 150 lines of the mpatrol*log file :).

Issue History

Date Modified Username Field Change
2004-01-06 00:57 crimson New Issue
2004-01-06 01:16 codemastr Note Added: 0004540
2004-01-06 13:43 syzop Note Added: 0004554
2004-01-06 13:43 syzop Note Edited: 0004554
2004-01-06 16:23 codemastr Note Added: 0004556
2004-01-06 17:57 crimson Note Added: 0004557
2004-01-06 17:58 crimson Note Edited: 0004557
2004-01-06 19:15 syzop Note View State: 4557: private
2004-01-06 19:18 syzop Note Added: 0004558
2004-01-06 19:22 crimson Note Added: 0004559
2004-01-06 19:28 syzop Note Edited: 0004558
2004-01-06 19:29 syzop Note View State: 4557: public
2004-01-06 19:31 syzop Note Added: 0004560
2004-01-06 19:32 syzop Note Edited: 0004560
2004-01-06 19:39 crimson Note Added: 0004561
2004-01-06 22:18 crimson Note Added: 0004568
2004-01-30 22:15 syzop Note Added: 0004802
2004-01-31 09:37 crimson Note Added: 0004808
2004-01-31 17:38 codemastr Note Added: 0004813
2004-01-31 17:47 syzop Note Added: 0004814
2004-01-31 18:20 codemastr Note Added: 0004815
2004-01-31 22:17 crimson Note Added: 0004825
2004-02-03 01:58 syzop Status new => closed
2004-02-03 01:58 syzop Note Added: 0004853