<LittleVulpix> nurupo: Oh? Let's see...
toxme.io is running so
oh wiw
wow*
[2152112.531883] Out of memory: Kill process 3782 (tox-bootstrapd) score 893 or sacrifice child
[2152112.643896] oom_reaper: reaped process 3782 (tox-bootstrapd), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB
If so many went offline at that time I would imagine someone caused a leak to manifest in tox bootstrap daemon took down several nodes with it that way
nurupo: node started but yeah I'm pretty sure this wouldn't have been random. I doubt we all started our nodes exactly at the same time so it must have been someone doing something with the network that caused bootstrap nodes to die
<nurupo> are you sure it was tox-bootstrapd that filled up your ram and not some other program?
<LittleVulpix> nurupo: not minding the fact the same happened to several completely unrelated machines, yeah I'm pretty sure
[2152112.531817] [ pid ] uid tgid total_vm rss nr_ptes nr_pmds swapents oom_score_adj name
[2152112.531846] [ 3782] 999 3782 467797 455738 903 4 0 0 tox-bootstrapd
467MB ram
reserved
455 actually used
normal resource usage is about 10-15MB RAM from what I remember
soooo that's like 45x as much
<iphy> LittleVulpix: interesting
<nurupo> if there is a memory leak then it's likely to be in toxcore itself so every tox node (i.e. even a client), i don't think there is anything in tox-bootstrapd that can leak memory, so everything that uses toxcore is likely affected, even clients. also, a lot of nodes were using old toxcore, like irungentoo old, so if there is a memory leak issue - it must be long present
<iphy> yes
The text was updated successfully, but these errors were encountered:
I have a core of my still running bootstrap node, which is up to 3.2GB res. Willing to work with someone to dig into it to root cause, ping me on Tox or IRC when a toxcore dev has time.
Happened on 2018-09-25.
Seems like there is a memleak in the code.
The text was updated successfully, but these errors were encountered: