|
Home > Archive > Unix administration > October 2005 > centralizing crontab for easier management
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
centralizing crontab for easier management
|
|
| Tom Van Overbeke 2005-10-24, 3:49 pm |
| Hi,
I manage a mix of solaris and linux servers. one of my main concerns for the
moment is that fact that we have horrible backup schedules (lots of
different applications require specialized backup scripts that create
files/archives that we need to put on tape afterwards). sometimes a job can
fail without us noticing because the error report didn't reach us via mail.
Due to the nature of the very diverse backups, we have no centralized backup
software (such as legato, netbackup or ....). Did I mention that we don't
have a budget for such software as well ?
What I would like to implement is some sort of a job scheduler that runs on
a server and pushes the jobs to run to the destination server (be it linux,
solaris or *nix), and reports back to the 'master' with both stdout and
stderr.
Now, I am aware of tools such as openpbs and sun's gridengine that - in
theory - could be configured to do this, but I was wondering if anyone has
ever used these tools for such purposes (they're mostly used for grids and
hpc).
or can so. suggest sth. better ?
thx,
tom.
| |
|
| Tom Van Overbeke wrote:
> Hi,
>
> I manage a mix of solaris and linux servers. one of my main concerns for the
> moment is that fact that we have horrible backup schedules (lots of
> different applications require specialized backup scripts that create
> files/archives that we need to put on tape afterwards). sometimes a job can
> fail without us noticing because the error report didn't reach us via mail.
> Due to the nature of the very diverse backups, we have no centralized backup
> software (such as legato, netbackup or ....). Did I mention that we don't
> have a budget for such software as well ?
>
> What I would like to implement is some sort of a job scheduler that runs on
> a server and pushes the jobs to run to the destination server (be it linux,
> solaris or *nix), and reports back to the 'master' with both stdout and
> stderr.
>
> Now, I am aware of tools such as openpbs and sun's gridengine that - in
> theory - could be configured to do this, but I was wondering if anyone has
> ever used these tools for such purposes (they're mostly used for grids and
> hpc).
>
> or can so. suggest sth. better ?
>
>
> thx,
>
> tom.
>
>
Rsync?
| |
| Dave Hinz 2005-10-24, 3:49 pm |
| On Mon, 10 Oct 2005 20:59:15 -0400, Ben <ben@en-ninguna-parte.com> wrote:
> Tom Van Overbeke wrote:
[vbcol=seagreen]
> Rsync?
There are environments where 'r' services haven't been turned off yet?
| |
| Michael Heiming 2005-10-24, 3:49 pm |
| In comp.unix.admin Dave Hinz <DaveHinz@spamcop.net>:
> On Mon, 10 Oct 2005 20:59:15 -0400, Ben <ben@en-ninguna-parte.com> wrote:
>
[vbcol=seagreen]
> There are environments where 'r' services haven't been turned off yet?
rsync -e ssh ....?
--
Michael Heiming (X-PGP-Sig > GPG-Key ID: EDD27B94)
mail: echo zvpunry@urvzvat.qr | PERL -pe 'y/a-z/n-za-m/'
#bofh excuse 14: sounds like a Windows problem, try calling
Microsoft support
| |
| Dave Hinz 2005-10-24, 3:49 pm |
| On Tue, 11 Oct 2005 16:42:43 +0200, Michael Heiming <michael+USENET@www.heiming.de> wrote:
> In comp.unix.admin Dave Hinz <DaveHinz@spamcop.net>:
[vbcol=seagreen]
[vbcol=seagreen]
[vbcol=seagreen]
> rsync -e ssh ....?
You know, right after I hit send, I had that "hey, idiot, you're using
rsync over ssh all over the place". Ah well...maybe I was thinking it
said "rdist" or something, who knows. That or it's too early/not enough
coffee yet. Yes, I think that is more likely.
| |
| Michael Heiming 2005-10-24, 3:49 pm |
| In comp.unix.admin Dave Hinz <DaveHinz@spamcop.net>:
> On Tue, 11 Oct 2005 16:42:43 +0200, Michael Heiming <michael+USENET@www.heiming.de> wrote:
>
[vbcol=seagreen]
[vbcol=seagreen]
[vbcol=seagreen]
> You know, right after I hit send, I had that "hey, idiot, you're using
> rsync over ssh all over the place". Ah well...maybe I was thinking it
> said "rdist" or something, who knows. That or it's too early/not enough
> coffee yet. Yes, I think that is more likely.
rdist -P /usr/bin/ssh ...?
Get some more coffee? ;-)
--
Michael Heiming (X-PGP-Sig > GPG-Key ID: EDD27B94)
mail: echo zvpunry@urvzvat.qr | PERL -pe 'y/a-z/n-za-m/'
#bofh excuse 312: incompatible bit-registration operators
| |
| Dave Hinz 2005-10-24, 3:49 pm |
| On Tue, 11 Oct 2005 17:06:18 +0200, Michael Heiming <michael+USENET@www.heiming.de> wrote:
> In comp.unix.admin Dave Hinz <DaveHinz@spamcop.net>:
[vbcol=seagreen]
> rdist -P /usr/bin/ssh ...?
Figures...
> Get some more coffee? ;-)
OK, did that, now what?
| |
| Stein Arne Storslett 2005-10-24, 3:49 pm |
| <tomvo@hahaha.be> wrote in <die191$ogf$1@pop-news.nl.colt.net>:
> Hi,
>
> I manage a mix of solaris and linux servers. one of my main concerns for the
> moment is that fact that we have horrible backup schedules (lots of
> different applications require specialized backup scripts that create
> files/archives that we need to put on tape afterwards). sometimes a job can
> fail without us noticing because the error report didn't reach us via mail.
> Due to the nature of the very diverse backups, we have no centralized backup
> software (such as legato, netbackup or ....). Did I mention that we don't
> have a budget for such software as well ?
>
> What I would like to implement is some sort of a job scheduler that runs on
> a server and pushes the jobs to run to the destination server (be it linux,
> solaris or *nix), and reports back to the 'master' with both stdout and
> stderr.
>
> Now, I am aware of tools such as openpbs and sun's gridengine that - in
> theory - could be configured to do this, but I was wondering if anyone has
> ever used these tools for such purposes (they're mostly used for grids and
> hpc).
>
> or can so. suggest sth. better ?
I would suggest you look at some sort of batch scheduling software
(Control-M from BMC, TWS from IBM, etc are examples).
That way you can centralize the scheduling and supervising of your jobs.
If you find any good OpenSource you could drop a line about it in this
group. Probably more ppl interested.
With a tight budget you could also make your backup scripts write a
short status to a log. You then can write a script to create an HTML-page
you check manually every day (so you won't have to depend on email).
The script could write an errormessage if it didn't find any status from
a given system.
--
Stein Arne
| |
| Simon Hooper 2005-10-24, 3:49 pm |
| Hi Tom,
On 2005-10-10, Tom Van Overbeke <tomvo@hahaha.be> wrote:
> I manage a mix of solaris and linux servers. one of my main concerns for the
> moment is that fact that we have horrible backup schedules (lots of
> different applications require specialized backup scripts that create
> files/archives that we need to put on tape afterwards). sometimes a job can
> fail without us noticing because the error report didn't reach us via mail.
> Due to the nature of the very diverse backups, we have no centralized backup
> software (such as legato, netbackup or ....). Did I mention that we don't
> have a budget for such software as well ?
Have you heard of Amanda? It's a free (GPL) network disk archiver.
www.amanda.org
HTH
Simon.
--
Simon Hooper
| |
| Dave Hinz 2005-10-24, 3:49 pm |
| On Wed, 12 Oct 2005 10:07:24 +0000 (UTC), Simon Hooper <simon@trickshot.nu> wrote:
> Hi Tom,
> Have you heard of Amanda? It's a free (GPL) network disk archiver.
> www.amanda.org
Does that help with central scheduling?
| |
| Brian Gutman 2005-10-24, 3:49 pm |
| If you use one machine running cron, you could write your shell scripts to
connect to the remote host and fire off the backup using whatever software
package you choose. I use this method in a production environment where I
have 3 groups of 30 machines each, and it works great, and have no problems
in last 4.5 yrs.
--
Brian Gutman
| |
| apairudin 2005-10-24, 3:49 pm |
| Cfengine is good for centralised management. We use it for Linux and
Solaris. Not only for your crontabs, but many other stuff (patch
management, centralised apps management, etc).
Cfengine.org
| |
| Kyle Tucker 2005-10-24, 3:49 pm |
| In article <1129320351.489307.128590@g43g2000cwa.googlegroups.com>,
"apairudin" <apairudin@gmail.com> writes:
> Cfengine is good for centralised management. We use it for Linux and
> Solaris. Not only for your crontabs, but many other stuff (patch
> management, centralised apps management, etc).
Do you use it for Solaris packages as well? I've been looking for
contributed cfengine configs for patches and packages and haven't
seen much published. I would be very interested in seeing what you
have designed if you'd be willing to make it available.
--
- Kyle
| |
|
| Dave Hinz wrote:
> On Tue, 11 Oct 2005 17:06:18 +0200, Michael Heiming <michael+USENET@www.heiming.de> wrote:
>
>
>
>
>
>
>
> Figures...
>
>
>
>
> OK, did that, now what?
>
Remove foot from mouth ...
| |
| Dave Hinz 2005-10-24, 3:49 pm |
| On Fri, 14 Oct 2005 20:45:23 -0400, Ben <ben@en-ninguna-parte.com> wrote:
[vbcol=seagreen]
> Remove foot from mouth ...
(shrug) Been wrong before, I suspect it'll happen again...
|
|
|
|
|