2010-12-16 00:03:54 utc |
InformatiQ |
hello |
2010-12-16 00:04:05 utc |
InformatiQ |
If i have a ruote instance |
2010-12-16 00:04:14 utc |
InformatiQ |
i launched a process |
2010-12-16 00:04:25 utc |
InformatiQ |
i made a mistake and it won't finish |
2010-12-16 00:04:39 utc |
InformatiQ |
how do i stop it? |
2010-12-16 00:07:36 utc |
jmettraux |
hello, welcome to #ruote |
2010-12-16 00:07:44 utc |
jmettraux |
do you have the wfid of the instance ? |
2010-12-16 00:07:50 utc |
jmettraux |
wfid = engine.launch(pdef) |
2010-12-16 00:08:03 utc |
jmettraux |
if you have it you can do engine.cancel(wfid) |
2010-12-16 00:08:47 utc |
jmettraux |
engine.cancel_process(wfid) |
2010-12-16 00:08:50 utc |
InformatiQ |
jmettraux: the problem i am not a "ruote" user but rather a "boss" user |
2010-12-16 00:09:13 utc |
InformatiQ |
so i am not sure where do i run that |
2010-12-16 00:09:17 utc |
jmettraux |
sorry, then could you please ask lbt (David Greaves) |
2010-12-16 00:09:30 utc |
brain_in_vat |
just turn the machine off and then on again |
2010-12-16 00:09:35 utc |
brain_in_vat |
(sorry) |
2010-12-16 00:09:36 utc |
jmettraux |
lol |
2010-12-16 00:09:37 utc |
InformatiQ |
jmettraux: yeah i thought i might ask here sinse he is sleeping |
2010-12-16 00:09:51 utc |
InformatiQ |
by the way I am ramez hanna from nokia |
2010-12-16 00:10:05 utc |
jmettraux |
Nice to meet you Ramez, I'm John |
2010-12-16 00:10:18 utc |
jmettraux |
and this is Nicholas |
2010-12-16 00:10:48 utc |
jmettraux |
InformatiQ: what does you system look like ? Web interface ? |
2010-12-16 00:10:55 utc |
InformatiQ |
I am the team lead for david and others |
2010-12-16 00:11:08 utc |
InformatiQ |
jmettraux: no web |
2010-12-16 00:11:26 utc |
InformatiQ |
jmettraux: are you asking about my current problem or in general |
2010-12-16 00:11:44 utc |
jmettraux |
about your current problem, trying to help you as best as I can |
2010-12-16 00:11:46 utc |
brain_in_vat |
hi InformatiQ |
2010-12-16 00:11:57 utc |
InformatiQ |
hi brain_in_vat |
2010-12-16 00:12:38 utc |
InformatiQ |
jmettraux: i launch processes using python |
2010-12-16 00:12:55 utc |
InformatiQ |
have several participants to do the work |
2010-12-16 00:13:10 utc |
InformatiQ |
brain_in_vat: would it really work if i reboot? |
2010-12-16 00:13:19 utc |
InformatiQ |
or maybe kill ruote and restart it |
2010-12-16 00:13:38 utc |
jmettraux |
then, in the same class that you use for the launch, there is probably a cancel or cancel_process method |
2010-12-16 00:13:46 utc |
brain_in_vat |
InformatiQ: I don't know, but I was joking |
2010-12-16 00:14:19 utc |
jmettraux |
killing and restarting ruote : the workflow instance would restart as soon as ruote is restarted |
2010-12-16 00:14:28 utc |
InformatiQ |
jmettraux: that sounds sensible, i'll see |
2010-12-16 00:15:22 utc |
jmettraux |
InformatiQ: I always forget : what's the link to the BOSS source code ? |
2010-12-16 00:17:08 utc |
InformatiQ |
http://meego.gitorious.org/meego-infrastructure-tools and docs at http://wiki.meego.com/Release_Infrastructure/ |
2010-12-16 00:17:49 utc |
InformatiQ |
oh cance is not yet implemented in the libs i am using |
2010-12-16 00:18:47 utc |
InformatiQ |
gotta talk to lbt about that as he implemeted it AFAIK |
2010-12-16 00:24:00 utc |
InformatiQ |
well ruote must store all those processes on disk |
2010-12-16 00:26:08 utc |
InformatiQ |
at the moment this is my test env. so i can destroy anything |
2010-12-16 00:27:47 utc |
jmettraux |
ah ues |
2010-12-16 00:27:50 utc |
jmettraux |
yes |
2010-12-16 00:28:19 utc |
InformatiQ |
now grepping for parts of that process |
2010-12-16 00:28:25 utc |
InformatiQ |
i beleive that could help |
2010-12-16 00:28:58 utc |
jmettraux |
you can nuke a process manually by destroyed all the files that have its wfid in their names |
2010-12-16 00:29:05 utc |
jmettraux |
destroying |
2010-12-16 00:29:28 utc |
InformatiQ |
good to know that |
2010-12-16 00:29:41 utc |
jmettraux |
find work -name "*20101216-bipemupushi*" | xargs rm |
2010-12-16 00:29:52 utc |
InformatiQ |
i'm a sadistic sys admin who likes to nuke systems |
2010-12-16 00:30:27 utc |
jmettraux |
given a "work" ruote work directory, this command would do it for a wfid of 20101216-bipemupushi |
2010-12-16 00:33:00 utc |
InformatiQ |
that seemed to work good |
2010-12-16 00:33:34 utc |
jmettraux |
:-) |
2010-12-16 00:34:42 utc |
InformatiQ |
jmettraux: thanks |
2010-12-16 00:34:48 utc |
jmettraux |
you're welcome |
2010-12-16 00:34:55 utc |
InformatiQ |
that was driving crazy sinse 3 hours ago |
2010-12-16 00:35:35 utc |
InformatiQ |
i run a process and i get a totally different behaviour that i had wrote earlier but expected it to have ended |
2010-12-16 00:39:46 utc |
jmettraux |
InformatiQ: another way of doing some looking/tinkering, if you have ruby and the right rubygems : https://gist.github.com/742848 |
2010-12-16 00:41:20 utc |
jmettraux |
and this one cancels the process instance whose wfid is given as first argument of the command line : https://gist.github.com/742852 |
2010-12-16 00:42:20 utc |
jmettraux |
these two scripts are just "consoles", the first one just behaves like a "ps" and the second is merely a "kill -9 wfid" |
2010-12-16 00:42:52 utc |
jmettraux |
the second one just passes the order to cancel, it assumes there is a ruote worker active that will do the cancelling work |
2010-12-16 00:43:32 utc |
InformatiQ |
this is very helpfull jmettraux i'm trying it now |
2010-12-16 00:43:47 utc |
InformatiQ |
cause i beleive i have hundreds of those stale ones |
2010-12-16 00:45:10 utc |
jmettraux |
if you want to start fresh, you can stop ruote, nuke its work directory and restart ruote |