Home My Page Projects cado-nfs
Summary Activity Forums Tracker Lists Tasks Docs News SCM Files

[#19261] cadofactor complains about clients it just launched in multiple groups

Date:
2015-06-25 14:26
Priority:
1
State:
Open
Submitted by:
Nadia Heninger (nadiah)
Assigned to:
Paul Zimmermann (zimmerma)
Hardware:
none
Product:
none
Operating System:
none
Component:
none
Version:
none
Severity:
none
Resolution:
Works For Me
URL:
Summary:
cadofactor complains about clients it just launched in multiple groups

Detailed description
When using the params file to specify multiple groups of clients to be launched, a warning message will be displayed after every group is launched complaining about all of the clients it just launched in the previous round.

Eg:
...
Info:Client Launcher: Starting client id lattice0+20 on host lattice0
...
Info:Client Launcher: Running clients: lattice0+20 (Host lattice0, PID 39176) ...
...
Warning:Client Launcher: Client id lattice0+20 (Host lattice0, PID 39176), launched in a previous run and not meant to be launched this time, is still running
Message  ↓
Date: 2015-11-12 10:54
Sender: Paul Zimmermann

I can now reproduce the issue with revision cec716b and the attached params.c90:

zimmerma@tarte:~/svn/cado-nfs$ ./cado-nfs.py /tmp/params.c90
...
Info:Client Launcher: Starting client id fondue on host fondue
Info:Client Launcher: Starting client id fondue+2 on host fondue
Info:Client Launcher: Running clients: fondue (Host fondue, PID 16783), fondue+2 (Host fondue, PID 16835)
Info:Client Launcher: Starting client id berthoud on host berthoud
Info:Client Launcher: Starting client id berthoud+2 on host berthoud
Info:Client Launcher: Running clients: berthoud (Host berthoud, PID 15547), fondue (Host fondue, PID 16783), fondue+2 (Host fondue, PID 16835), berthoud+2 (Host berthoud, PID 15596)
Warning:Client Launcher: Client id fondue (Host fondue, PID 16783), launched in a previous run and not meant to be launched this time, is still running
Warning:Client Launcher: Client id fondue+2 (Host fondue, PID 16835), launched in a previous run and not meant to be launched this time, is still running
Info:Client Launcher: Starting client id sel on host sel
Info:Client Launcher: Starting client id sel+2 on host sel
Info:Client Launcher: Running clients: berthoud (Host berthoud, PID 15547), berthoud+2 (Host berthoud, PID 15596), fondue+2 (Host fondue, PID 16835), fondue (Host fondue, PID 16783), sel+2 (Host sel, PID 6500), sel (Host sel, PID 6452)
Warning:Client Launcher: Client id berthoud (Host berthoud, PID 15547), launched in a previous run and not meant to be launched this time, is still running
Warning:Client Launcher: Client id fondue (Host fondue, PID 16783), launched in a previous run and not meant to be launched this time, is still running
Warning:Client Launcher: Client id fondue+2 (Host fondue, PID 16835), launched in a previous run and not meant to be launched this time, is still running
Warning:Client Launcher: Client id berthoud+2 (Host berthoud, PID 15596), launched in a previous run and not meant to be launched this time, is still running

Date: 2015-11-10 08:16
Sender: Paul Zimmermann

works for me with the following in params.c90 and revision 26890db:

slaves.big.hostnames = fondue,berthoud
slaves.big.nrclients = 4
slaves.big.scriptpath = /users/caramel/zimmerma/svn/cado-nfs
slaves.big.basepath = /tmp/nfs
slaves.small.hostnames = sel,poivre
slaves.small.nrclients = 2
slaves.small.scriptpath = /users/caramel/zimmerma/svn/cado-nfs
slaves.small.basepath = /tmp/nfs

Date: 2015-11-10 08:14
Sender: Paul Zimmermann

from Nadia:

Here's an example of the relevant part of the config file for
launching multiple groups. The behavior I was seeing if I remember
correctly was that it would launch, say, the first group, and then try
to relaunch each of them when launching the second group.

slaves.big.hostnames = cx0,qubit0,qubit1,lattice0
slaves.big.nrclients = 24
slaves.big.scriptpath = /home/nadiah/cado-nfs/scripts/cadofactor
slaves.big.basepath = /home/nadiah/workdir
slaves.small.hostnames = cx1,cx2,cx3,cx4,cx5
slaves.small.nrclients = 8
slaves.small.scriptpath = /home/nadiah/cado-nfs/scripts/cadofactor
slaves.small.basepath = /home/nadiah/workdir

Attachments:
Size Name Date By Download
22 KiBparams.c902015-11-12 10:54zimmermaparams.c90
Field Old Value Date By
File Added5124: params.c902015-11-12 10:54zimmerma
assigned_tonone2015-11-10 08:16zimmerma
ResolutionNone2015-11-10 08:16zimmerma
summarycadofactor complains about clients it just launched imultiple groups2015-06-25 14:29nadiah