Opened 12 years ago

Last modified 7 years ago

#404 closed defect (Fixed)

seg fault on 0.1.2.9-rc dir authority

Reported by: arma Owned by:
Priority: Low Milestone:
Component: Core Tor/Tor Version: 0.1.2.10-rc
Severity: Keywords:
Cc: arma, nickm Actual Points:
Parent ID: Points:
Reviewer: Sponsor:

Description

Mar 10 09:52:01.416 [info] dirserv_orconn_tls_done(): Found router liquidvibration3 to be reachable. Yay.
Mar 10 09:52:01.542 [info] routerlist_remove_old_routers(): Forgetting obsolete (too old) routerinfo for router 'mrs5'
Mar 10 09:52:01.562 [info] dirserv_pick_cached_dir_obj(): The server directory is still clean; reusing.
Mar 10 09:52:01.711 [err] routerlist.c:4485: routerlist_assert_ok: Assertion r == r2 failed; aborting.

This was moria1. I'll hopefully get a core if it happens a second time.

[Automatically added by flyspray2trac: Operating System: All]

Child Tickets

Change History (9)

comment:1 Changed 12 years ago by nickm

I checked some more asserts into svn trunk this afternoon to try to trigger this earlier.

comment:2 Changed 12 years ago by nickm

I've re-opened bug 404: It looks like the bug that could have caused 417 (and all its friends) didn't exist in 0.1.2.x.

comment:3 Changed 12 years ago by nickm

It is very likely that the underlying causes of this are fixed in svn trunk.

To debug this more, though, we'll need an authority op who tries running the 0.1.2.x branch on an authority,
and who's willing to send in backtraces.

Any takers?

comment:4 Changed 12 years ago by arma

Lefkada just triggered this bug this evening, running 0.1.2.14.

His backtrace didn't have any symbols in it though Perhaps weasel can
walk goodell through installing the package that has the symbols?

comment:5 Changed 12 years ago by arma

Patch that looks related to this: r10217

comment:6 Changed 12 years ago by goodell

It appears that this bug may have caused lefkada to crash today around 19:09 GMT.

May 29 16:09:07.300 [err] routerlist.c:4539: routerlist_assert_ok: Assertion r == r2 failed; aborting.

...backtrace:

21:42 goodell - #0 0xffffe410 in kernel_vsyscall ()
21:42 goodell - #1 0xb7c57811 in raise () from /lib/tls/i686/cmov/libc.so.6
21:42 goodell - #2 0xb7c58fb9 in abort () from /lib/tls/i686/cmov/libc.so.6
21:42 goodell - #3 0x080ab012 in routerlist_assert_ok (rl=0x11bb) at routerlist.c:4531
21:42 goodell - #4 0x080acb78 in routerlist_remove_old_routers () at routerlist.c:2152
21:42 goodell - #5 0x080acf07 in router_have_minimum_dir_info () at routerlist.c:4262
21:42 goodell - #6 0x08094703 in second_elapsed_callback (fd=-1, event=1, args=0x0) at main.c:932
21:42 goodell - #7 0xb7d62c79 in event_base_priority_init () from /usr/lib/libevent-1.1a.so.1
21:42 goodell - #8 0xb7d62f65 in event_base_loop () from /usr/lib/libevent-1.1a.so.1
21:42 goodell - #9 0xb7d62dcb in event_loop () from /usr/lib/libevent-1.1a.so.1
21:42 goodell - #10 0xb7d62cb0 in event_dispatch () from /usr/lib/libevent-1.1a.so.1
21:42 goodell - #11 0x08095957 in tor_main (argc=1, argv=0xbfc29a64) at main.c:1267
21:42 goodell - #12 0x080b93e2 in main (argc=Cannot access memory at address 0x2c1a

...more:

21:51 goodell - #4 0x080acb78 in routerlist_remove_old_routers () at routerlist.c:2152
21:51 goodell - 2152 in routerlist.c
21:51 goodell - (gdb) p routerlist
21:51 goodell - $10 = (routerlist_t *) 0x813f7f0
21:51 goodell - (gdb) p *routerlist
21:51 goodell - $11 = {identity_map = 0x8141c48, desc_digest_map = 0x8142858, routers = 0x8143e80,
21:51 goodell - old_routers = 0x8144cb8, mmap_descriptors = 0x8144ca8}
21:51 goodell - (gdb) p networkstatus_list
21:51 goodell - $12 = (smartlist_t *) 0x8142a58
21:51 goodell - (gdb) p *networkstatus_list
21:51 goodell - $13 = {list = 0x813f6b8, num_used = 5, capacity = 32}
21:51 goodell - (gdb) p now
21:51 goodell - $14 = 1180469347
21:51 goodell - (gdb) p cutoff
21:51 goodell - $15 = <value optimized out>

comment:7 Changed 12 years ago by nickm

01:15 < nickm> goodell: if your core is still around, I'd really like to know

what r and r2 and r_sl_idx are in assert_routerlist_ok().

01:18 < armadev> nickm: you should put that request on the 404 comments list too

comment:8 Changed 12 years ago by weasel

flyspray2trac: bug closed.
let's hope r1075[234] fixed this one

comment:9 Changed 7 years ago by nickm

Component: Tor RelayTor
Note: See TracTickets for help on using tickets.