00:28
supersandro2000 has quit [Disconnected by services]
00:28
supersandro2000 has joined #nixos-security
01:03
{^_^} has joined #nixos-security
01:07
star_cloud has quit [Ping timeout: 268 seconds]
01:08
justanotheruser has quit [Ping timeout: 268 seconds]
01:19
star_cloud has joined #nixos-security
01:21
rajivr has joined #nixos-security
01:26
star_cloud has quit [Ping timeout: 240 seconds]
01:28
star_cloud has joined #nixos-security
01:38
star_cloud has quit [Excess Flood]
01:39
star_cloud has joined #nixos-security
02:10
gchristensen has joined #nixos-security
03:08
glowpelt has quit [Max SendQ exceeded]
03:09
glowpelt has joined #nixos-security
05:15
justanotheruser has joined #nixos-security
06:24
cole-h has quit [Ping timeout: 264 seconds]
07:27
anselmolsm has quit [Remote host closed the connection]
07:29
anselmolsm has joined #nixos-security
08:54
anselmolsm has quit [Remote host closed the connection]
08:58
anselmolsm has joined #nixos-security
09:41
anselmolsm_ has joined #nixos-security
09:41
anselmolsm has quit [Ping timeout: 268 seconds]
09:45
anselmolsm_ has quit [Remote host closed the connection]
09:47
anselmolsm_ has joined #nixos-security
12:02
star_cloud has quit [Ping timeout: 260 seconds]
13:40
star_cloud has joined #nixos-security
15:56
red[evilred] has joined #nixos-security
15:56
<
red[evilred] >
something just caused systemd to SEGV
15:58
<
red[evilred] >
that's new
15:59
<
red[evilred] >
nice - now I can't reboot:
16:04
<
red[evilred] >
Okay - so this is interesting...
16:05
<
red[evilred] >
someone connected on my high ssh port
16:05
<
red[evilred] >
and apparently caused systemd to crash
16:09
__red__ has quit [Ping timeout: 256 seconds]
16:09
red[evilred] has quit [Ping timeout: 264 seconds]
16:10
bridge[evilred] has quit [Ping timeout: 264 seconds]
16:14
bridge[evilred] has joined #nixos-security
16:15
__red__ has joined #nixos-security
16:16
<
__red__ >
so - I don't know for sure if my matrix server passed this on or not
16:16
<
__red__ >
but apparently - someone connected to my sshd server and caused systemd to SEGV
16:17
<
tilpner >
Yes, up until "and apparently caused systemd to crash"
16:17
<
__red__ >
it was my matrix node, so I didn't know if it made it or if anyone replied
16:17
<
__red__ >
thank you
16:18
<
__red__ >
Gonna assume neutrino unless I see it again
16:18
anselmolsm has joined #nixos-security
16:18
anselmolsm_ has quit [Ping timeout: 240 seconds]
16:18
<
__red__ >
but it's a first for me for sure
16:18
<
tilpner >
Yeah, it looks very odd. So no ECC then?
16:19
red[evilred] has joined #nixos-security
16:19
<
red[evilred] >
it's a digital ocean VM
16:19
<
red[evilred] >
so I would have hoped so
16:19
<
red[evilred] >
the timing is what worries me
16:29
<
tilpner >
red[evilred]: Apparently that's a user-mode pagefault, so a bitflip in a pointer might do it. But DO says they use ECC. Unless you have that coredump, maybe run online memcheck?
16:31
<
red[evilred] >
where would systemd put its core files? / ?
16:33
<
tilpner >
I'm not sure what the errors mean, the whole "dumping: No such process" thing is confusing me
16:34
<
tilpner >
ls -t /var/lib/systemd/coredumps
16:34
<
red[evilred] >
yup - got one
16:35
<
tilpner >
Or find / -name \*32640\*
16:35
<
red[evilred] >
core.systemd.0.832d7e879c8d4a5d9ca6a8e0221a8da4.32640.1608218689000000.lz4: LZ4 compressed data (v1.4+)
16:35
<
red[evilred] >
so, uncompress it
16:36
<
tilpner >
Now there's the only the hard part left, making sense of it \o/
16:36
<
red[evilred] >
gdb /path/to/systemd /path/to/coredump
16:36
<
red[evilred] >
and backtrace?
16:36
<
red[evilred] >
(it's been a while)
16:37
<
red[evilred] >
well, looks like I have my morning project :-)
16:37
<
tilpner >
Yeah, something like that
16:40
<
red[evilred] >
I think that core file is a trap:
16:40
<
red[evilred] >
core.systemd.0.832d7e879c8d4a5d9ca6a8e0221a8da4.32640.1608218689000000.lz4
16:40
<
red[evilred] >
Dictionary size: 1026 MB (2^30 bytes)
16:40
<
red[evilred] >
Uncompressed size: 8762807027734 MB (9188469141913302864 bytes)
16:41
<
red[evilred] >
need a bigger disk
16:41
<
red[evilred] >
8.7PB core file if my math is correct
16:41
<
tilpner >
Does coredumpctl debug do anything else?
16:42
<
red[evilred] >
oh - never heard of that
16:42
<
red[evilred] >
ah - that worked
16:44
<
red[evilred] >
yeah - don't know I can do anything with that
16:44
<
tilpner >
thread apply all bt
16:44
<
tilpner >
(in case it has multiple threads. I've never tried that with coredumps)
16:46
cole-h has joined #nixos-security
17:53
rajivr has quit [Quit: Connection closed for inactivity]
19:21
star_cloud has quit [Ping timeout: 268 seconds]
19:27
star_cloud has joined #nixos-security
19:46
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
20:15
cole-h has quit [Quit: Goodbye]
21:10
star_cloud has quit [Ping timeout: 240 seconds]
23:18
star_cloud has joined #nixos-security