Quoting Yonah Russ, from the post of Thu, 27 Apr:
I think everything suggested so far is way to complicated.
Have all your servers mount a directory(lets call it cronserver) and store
all your scheduled scripts there.
ok, talk about complicated. NFS is the LAST answer I want to give...
here's
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Yedidyah Bar-David wrote:
We used for some years, in cs.tau, openpbs. It mostly worked, but for
some reasons we were not happy with it. A few years ago we moved to
condor. I did not work with it much, so I don't have an opinion. It's
rather big
Hi Linuxers,
For large production system, Im looking for a scheduler
with the following capabilities:
Ability to run tasks on many
servers from one central location (Preferably using SSH).
Ability to process return code
from task after running it and show if it succeeded
I'd use an homebrew system, which employs a powerful scripting language.
First, design a data structure, which can describe all those operations
which you want to perform.
For each task, have task ID, the command to be executed, command to test
success/failure, command to execute in case of
On Wed, Apr 26, 2006 at 09:28:56AM +0300, Chen Shapira wrote:
Hi Linuxers,
For large production system, I'm looking for a scheduler with the
following capabilities:
* Ability to run tasks on many servers from one central location
(Preferably using SSH).
* Ability to process
be a su-chef if its
still needed.
-Original Message-
From: marc [mailto:[EMAIL PROTECTED]
Sent: Wednesday, April 26, 2006 10:38 AM
To: Yedidyah Bar-David
Cc: Chen Shapira; linux-il@cs.huji.ac.il
Subject: Re: Need recommendation for task scheduler
-BEGIN PGP SIGNED MESSAGE-
Hash
Chen Shapira [EMAIL PROTECTED] writes:
I probably didn't explain myself well, because I seem to get
suggestions that are a bit different from what I had in mind.
Ah, that's because the words schedule and scheduler are so heavily
overloaded... Sorry for the noise...
The idea is that I have
Oleg Goldshmidt wrote:
Chen Shapira [EMAIL PROTECTED] writes:
The idea is that I have jobs that I want every machine in my network
to do on a schedule: backups, exports, generate daily reports,
delete garbage files, run a test, etc. etc.
So why don't you just use cron on a central
Shachar Shemesh [EMAIL PROTECTED] writes:
You can even rig up a cron on a central management station that will
ssh to other machines and run the jobs via at(1).
If you're using cron already, why not run the entire process by
cron?
I believe this is true: ssh will tie local streams with the
On Wed, 26 Apr 2006, Chen Shapira wrote:
Hi,
I probably didn't explain myself well, because I seem to get suggestions
that are a bit different from what I had in mind.
The idea is that I have jobs that I want every machine in my network to
do on a schedule: backups, exports, generate daily
Orna Agmon wrote:
On Wed, 26 Apr 2006, Chen Shapira wrote:
For this I
would indeed combine it with a script which will deal with make's
failures, and perform a different target. Or is there a way to do this
from inside make?
Make rules are being processed by /bin/sh, seperate shell
the revival. I can also be a su-chef if its
still needed.-Original Message-From: marc [mailto:[EMAIL PROTECTED]]Sent: Wednesday, April 26, 2006 10:38 AMTo: Yedidyah Bar-DavidCc: Chen Shapira;
linux-il@cs.huji.ac.ilSubject: Re: Need recommendation for task scheduler-BEGIN PGP SIGNED
12 matches
Mail list logo