Subj : binkd crashes when reloading after file change detection
To : Michiel van der Vlist
From : mark lewis
Date : Thu Jan 20 2022 09:12 am
On 2022 Jan 20 08:55:17, I wrote to you:
MV>> I gave up and stopped using the -C option.
ml> i've continued to use it but have always stopped my binkd manually before
ml> making and saving edits... [...]
earlier i had remove the "-C" to test the automatic reload that the FAQ seems to indicate is done with binkd v1.xx... i've just put it back and run a short but similar test to your's using touch on my main configuration file as well as each of the various included files... in every case, binkd did detect the timestamp updates and said it reloaded the configure for both server and client instances... there was no crash...
ok, so that works... let's try changing some content...
first test is manually changing the case of a comment in my main binkd.conf file... both instances reloaded just fine... tested this several times changing only the case of some comment text... no problems... hummm...
second test is manually changing the case of a comment in the first include file, my binkd-networks.conf... again, both instances reloaded just fine... tested this several times, too... no problems... hummm...
third test is manually changing the case of a comment in another include file... same as the two previous tests... no crash... ok... hummm...
4th test... now we'll add a blank line to the end of the main config file... no crash... ok, remove that blank line... no crash... alright...
5th test... just like 4th, we'll add a blank line to the end of the first included file... no crash... ok... remove that blank line... and again, no crash...
this is getting curiousier and curiouser...
6th test... we'll change the data somewhere... we'll change "timeout 1m" to "timeout 45s" in the main conf file... that'll add one byte to the file, itself...no crash... ok, change it back... again, no crash...
7th test... now we'll change a data line in the first include file... that's the file that triggered my initial post in this thread... we'll comment out a domain line for a network known to no longer exist... all we're doing is adding a '#' to the beginning of that domain line... that'll add one byte and remove some data when binkd processes the contents... damn! again no crash... remove the content so that domain line is valid again and no crash... WTAF is going on??
8th test... instead of commenting out that domain line, i've completely removed it... no crash... put it back by copying a backup copy of the file back over the one we've edited... again no crash...
i don't get it...
[time passes]
ahhh... if i'm reading the output of "ps aux" properly, there is a bit of a memory leak going on... perhaps the crashes i was able to trigger are the result of a combination of some memory leakage over a "long" running time coupled with changing a config or included file? i'll try to set something up to test this possibility further but i'm done manually beating on it by editing files for now ;)
remove domain line as in 8th test
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
sbbs 17629 0.1 0.3 14648 3684 pts/1 S+ 09:48 0:00 binkd: server manager (listen 24554)
sbbs 17630 0.1 0.2 14708 2988 pts/1 S+ 09:48 0:00 binkd: client manager
copy back file over edited file restoring domain line as in 8th test
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
sbbs 17629 0.1 0.3 14784 3844 pts/1 S+ 09:48 0:00 binkd: server manager (listen 24554)
sbbs 17630 0.2 0.3 14868 3204 pts/1 S+ 09:48 0:00 binkd: client manager
the above may also not be quite scientific as there was at least one mail connection at one stage... i don't recall if there was any actual files moved, though... so i've whipped up a quick script that will monitor and log binkd's memory usage... it'll take a reading as above once every 5 minutes... let's see what we find...
where's my asprin?
)\/(ark
"The soul of a small kitten in the body of a mighty dragon. Look on my majesty, ye mighty, and despair! Or bring me catnip. Your choice. Oooh, a shiny thing!"
... There is no such thing as arugula yet many gourmet recipes call for it.
---
* Origin: (1:3634/12.73)
