Hello,
boinc-manager is configured not to leave apps in memory if not running:

$ grep leave_apps_in_memory /etc/boinc-client/global_prefs_override.xml
   <leave_apps_in_memory>0</leave_apps_in_memory>

but they are left there:

# ps aux | grep boinc
morph     1708  0.9  0.2  85408 10840 pts/7    S+   16:05   0:00 /usr/bin/python /usr/bin/reportbug -O boinc-manager
morph     4502  0.8  0.5 164024 21660 ?        Ss   Jul10  21:44 /usr/bin/boincmgr
boinc    24879 86.0  2.0 248300 82984 ?        SNl  12:50 168:41 Amolqc-preRC1-501 qmc
boinc    24959 85.7  2.0 248300 82472 ?        SNl  12:53 165:06 Amolqc-preRC1-501 qmc
boinc    29582  0.0  0.3  80560 14596 ?        SN   Jul10   2:22 /usr/bin/boinc --check_all_logins --redirectio --dir /var/lib/boinc-client
boinc    31086 10.4  9.3 1012356 381392 ?      SNl  15:02   6:41 garli_5.15_i686-pc-linux-gnu garli1686.conf
boinc    31090 10.1 10.0 1012360 408840 ?      SNl  15:02   6:29 garli_5.15_i686-pc-linux-gnu garli1698.conf
boinc    31290 79.4  1.6  79868 65604 ?        SNl  15:11  43:33 setiathome-5.28.x86_64-pc-linux-gnu
boinc    31291 81.4  1.0  56088 43156 ?        SNl  15:11  44:40 setiathome-5.28.x86_64-pc-linux-gnu

while only PID 24879 24959 31290 31291 are really running (other
processes are boinc itself), and I see the 2 garli processes in "Waiting
to run" state on boincmgr.

Regards,
Sandro

reassign 490543 boinc-client 6.2.11-1
tags 490543 unreproducible
thanks

Hi Sandro,

It is the BOINC client that controls applications, so this is an issue with
the client and not with the manager. I'm therefore reassigning this bug to
boinc-client.
[...]
and resuming them so that their status changed between "Running", "Task
suspended by user", and "Waiting to run". But ps indicated that only those
tasks with status "Running" were actually running.

Is this issue reproducible with your client and apps? Happens this only with
the garli applications or have you seen other applications that stayed in
memory although they were not "Running"? If you can reproduce this, could you
please set <task_debug> in /etc/boinc-client/cc_config.xml to 1 and post the
relevant output from the client's log file? Thanks!

BTW: In the first paragraph of this page[1] it is said that some applications
take fairly long to shut down. Maybe garli is one of those applications?

1:
http://www.boinc-wiki.info/Failed_to_stop_applications;_aborting_CPU_benchmarks

Cheers,
Frank

Hi Frank,

Yeah, sorry for the confusion ;)

I had 4 einstein processes running, I suspended 2 of them, and 2
abc-finder started, but 1 einstein is left in memory (the first one,
in sleep status)

# ps -eo pid,user,stat,comm | grep boinc
14097 boinc    SNl  einstein_S5R4_6
14100 boinc    RNl  abc-finder_1.03
14110 boinc    RNl  abc-finder_1.03
14150 morph    Ss   boincmgr
16603 boinc    SN   boinc
20173 boinc    RNl  einstein_S5R4_6
25929 boinc    RNl  einstein_S5R4_6

Configuration key changed, and just after restart, everything seems pretty ok:

# ps -eo pid,user,stat,comm | grep boinc
14459 boinc    SN   boinc
14485 boinc    RNl  einstein_S5R4_6
14488 boinc    RNl  einstein_S5R4_6
14489 boinc    RNl  abc-finder_1.03
14490 boinc    RNl  abc-finder_1.03
14497 morph    Ss   boincmgr

Then, keep suspending and resuming, I got:

# ps -eo pid,user,stat,comm | grep boinc
14459 boinc    SN   boinc
14485 boinc    RNl  einstein_S5R4_6
14497 morph    Ss   boincmgr
14525 boinc    RNl  abc-finder_1.03
14526 boinc    RNl  abc-finder_1.03
14570 boinc    SNl  abc-finder_1.03
14583 boinc    RNl  einstein_S5R4_6

I'm attaching the log generated with

grep "25-Aug-2008 22:5" /var/lib/boinc-client/stdoutdae.txt

yeah, here it's matter of minuts, for garli was some hours (then I
realized there was something that occupied all my memories and I've
restarted the boinc client).

Cheers,
Sandro