Double count of jobs

[D4rk4]

$ plotman status
plot id k tmp dst wall phase tmp pid stat mem user sys io
fc950a1b 32 /srv/chia/fast /srv/chia/dest 6:58 3:1 236G 4176527 SLP 2.6M 0s 0s 0s
fc950a1b 32 /srv/chia/fast /srv/chia/dest 6:58 3:1 236G 4176528 DSK 18.4G 6:01 0:07 1:23
a9f3ea7a 32 /srv/chia/fast /srv/chia/dest 9:40 4:0 176G 1317240 SLP 2.6M 0s 0s 0s
a9f3ea7a 32 /srv/chia/fast /srv/chia/dest 9:40 4:0 176G 1317241 RUN 17.7G 8:16 0:13 4:19

Chia from deb package:

$ dpkg -l | grep chia; which chia; ls -lash which chia
ii chia-blockchain 1.1.4 amd64 GUI for Chia Blockchain
/usr/local/bin/chia
512 lrwxrwxrwx 1 root staff 64 мая 8 17:57 /usr/local/bin/chia -> /usr/lib/chia-blockchain/resources/app.asar.unpacked/daemon/chia

[D4rk4]

#330

[altendky]

What is the output of ps aux | grep -P '(plots create|plotman)'? Also, how did you install plotman? Specifically @main or @development?

[jujuforce]

I have the exact same issue, I tired the command @altendky : it does return the job twice :(

[D4rk4]

ps aux | grep -P '(plots create|plotman)'

$ ps aux | grep -P '(plots create|plotman)' | grep -v grep
dark       31180  0.0  0.0   2536   612 ?        SNs  15:17   0:00 chia plots create -k 32 -r 6 -u 128 -b 8000 -t /srv/chia/fast -d /srv/chia/dest -f 82eba39f6428ca342fa0ce685f65f2dddb7441eeba1e23b861a96f57e8484b1500c69240a47fa4d2988aed32a8368b50 -p 9790ca1e0435c18839bdad5b0d4ef8cff762ab574549bcdd6852a83aa7ed9ad6ce79924dcacad1f77b60ad7ca654fa45
dark       31181  236  5.6 8953160 3742448 ?     SNLl 15:17 100:23 chia plots create -k 32 -r 6 -u 128 -b 8000 -t /srv/chia/fast -d /srv/chia/dest -f 82eba39f6428ca342fa0ce685f65f2dddb7441eeba1e23b861a96f57e8484b1500c69240a47fa4d2988aed32a8368b50 -p 9790ca1e0435c18839bdad5b0d4ef8cff762ab574549bcdd6852a83aa7ed9ad6ce79924dcacad1f77b60ad7ca654fa45
dark      229569  3.3  0.0  43116 24916 pts/0    S+   15:49   0:20 /usr/bin/python3 /home/dark/.local/bin/plotman interactive
dark      276992  0.0  0.0   2536   620 ?        SNs  15:59   0:00 chia plots create -k 32 -r 6 -u 128 -b 8000 -t /srv/chia/ext -d /srv/chia/dest -f 82eba39f6428ca342fa0ce685f65f2dddb7441eeba1e23b861a96f57e8484b1500c69240a47fa4d2988aed32a8368b50 -p 9790ca1e0435c18839bdad5b0d4ef8cff762ab574549bcdd6852a83aa7ed9ad6ce79924dcacad1f77b60ad7ca654fa45
dark      276993 67.7  0.2 706824 158988 ?       SNLl 15:59   0:19 chia plots create -k 32 -r 6 -u 128 -b 8000 -t /srv/chia/ext -d /srv/chia/dest -f 82eba39f6428ca342fa0ce685f65f2dddb7441eeba1e23b861a96f57e8484b1500c69240a47fa4d2988aed32a8368b50 -p 9790ca1e0435c18839bdad5b0d4ef8cff762ab574549bcdd6852a83aa7ed9ad6ce79924dcacad1f77b60ad7ca654fa45

It's @development branch

[altendky]

Thanks. I'll try to figure this out tonight. I do see #330 but would like to understand this a bit better. I'm a bit confused why this is just popping up now.

[jujuforce]

I got a fix on my computer but I won't create a pull request for this as this is a hack at best.
It is adding this filter in the get_running_jobs function :

proc.cpu_times().user == 0.0

It looks counter intuitive, but only the proc with this attributes at 0.0 does have open_files associated with it. I am not a linux nor python expert so I have no idea why ...

It works here but I guess that will not be the case on computer not affected by the issue.
A cleaner way would be to use the ppid() of the proc, but I can't make it work atm.

Cheers.

[lexzz]

I'm seeing the same issue on the development branch

[rafaelsteil]

I got a fix on my computer but I won't create a pull request for this as this is a hack at best.
It is adding this filter in the get_running_jobs function :

proc.cpu_times().user == 0.0

It looks counter intuitive, but only the proc with this attributes at 0.0 does have open_files associated with it. I am not a linux nor python expert so I have no idea why ...

I'd say it was more of a coincidence than anything else.

It works here but I guess that will not be the case on computer not affected by the issue.
A cleaner way would be to use the ppid() of the proc, but I can't make it work atm.

Relying on the parent pid can get tricky very quickly and doesn't indicate anything bad or good per se. A job started by plotman plot will have it's pid as parent as long as plotman plot is running. If you stop the plotter (a perfect valid scenario) the jobs will then return 1 as their parent PID.

[rafaelsteil]

$ plotman status
plot id k tmp dst wall phase tmp pid stat mem user sys io
fc950a1b 32 /srv/chia/fast /srv/chia/dest 6:58 3:1 236G 4176527 SLP 2.6M 0s 0s 0s
fc950a1b 32 /srv/chia/fast /srv/chia/dest 6:58 3:1 236G 4176528 DSK 18.4G 6:01 0:07 1:23
a9f3ea7a 32 /srv/chia/fast /srv/chia/dest 9:40 4:0 176G 1317240 SLP 2.6M 0s 0s 0s
a9f3ea7a 32 /srv/chia/fast /srv/chia/dest 9:40 4:0 176G 1317241 RUN 17.7G 8:16 0:13 4:19

@D4rk4 One thing that I noticed is that, despite that some columns are the same, other's aren't, which is curious. Can you confirm that when you saw this issue you actually had only two "real" jobs instead of 4?

[D4rk4]

@rafaelsteil it's 2 jobs, correct job info displayed in the second row of each plot ID

[altendky]

If you could try out #418, that'd be great. Follow the regular readme installation but use git+https://github.com/altendky/plotman@skip_matching_plot_process_parents as the url.

[jujuforce]

@altendky Works well for me, found a minor bug though, some temporary file (2-3 each time) are not removed after a plotman kill.

[altendky]

@juliencampy yeah, the plotter doesn't keep all the files open all the time I think and that's how we detect. There's some improvement to be done there for sure.

[D4rk4]

Work fine for me with pip install --force-reinstall git+https://github.com/ericaltendorf/plotman@development