Re: [Beaker-devel] Fwd: Re: take button in beaker
by Raymond Mancy
----- "Vladimir Benes" <vbenes(a)redhat.com> wrote:
> From: "Vladimir Benes" <vbenes(a)redhat.com>
> To: beaker-devel(a)lists.fedorahosted.org
> Sent: Thursday, September 23, 2010 10:49:19 PM GMT +10:00 Brisbane
> Subject: [Beaker-devel] Fwd: Re: take button in beaker
>
> resending from qe-dept list
>
>
> ----- "Petr Benas" <pbenas(a)redhat.com> wrote:
> >
> >
> > ----- "Vladimir Benes" <vbenes(a)redhat.com> wrote:
> > > Hi all,
> > > we discussed yesterday the possibility of returning take button into beaker. Maybe it can help something but as for now I use it like this.
> > >
> > > find myself any machine.. select the distro I want to provision and then click schedule provision (in the certain system page). This creates me a job and send me notification about it. So it works as well as take and even better in a way of scheduling currently occupied system as you don't have to check if it's already free.
> > > The job is not cancelled when rebooting so you can make whatever you want with the box.
> > >
> > > I don't see any reason for changing the code to get the Take button back, any ideas what went wrong so the need to have Take button came out?
> >
> > There is one, probably not sufficient enough since it's not too common case. Manual provisioning. If I want to install machine manually through vnc, need to perform it when the job starts. With oldschool provision I would reserve the machine and then reprovision it. This is now impossible. I reserve the machine, enqueue job with vnc install and return2beaker with hope no other jobs are in queue.
> >
> > I know I could use my own kickstart, but it seems to me it would cost me much more time. Actually, I don't need Take button, but ability to reprovision the machine.
>
> Just an idea, what about "Schedule provision" for machines where you're not current user and oldschool "Provision" if you're current user?
> This would meet needs Jan and I were describing and the infinite time reservation would remain deprecated, as was intention of Take button removal. Would this be hard to implement? Reprovisioning the machine and starting new reservesys with the time remaining...
I thought about doing this. The problem is that as soon as you go outside of the scheduler you are no longer beholden to any time restraints. Currently we have no automated way of nicely
booting a user from a machine (other than the scheduler/watchdog). At least for the moment, implementing this kind of feature is not a high priority for us AFAIK.
Cheers
Raymond
>
> >
> > >
> > > cheers,
> > > Vladimir
> > >
> >
> > --
> > Petr Benas
> > Kernel QA Brno
> > RedHat Czech, s.r.o
> > pbenas on #brno, #kernel-qe, #qa
> >
>
> --
> Petr Benas
> Kernel QA Brno
> RedHat Czech, s.r.o
> pbenas on #brno, #kernel-qe, #qa
>
> _______________________________________________ Beaker-devel mailing list Beaker-devel(a)lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/beaker-devel
13 years, 6 months
Fwd: Re: take button in beaker
by Vladimir Benes
resending from qe-dept list
----- "Petr Benas" <pbenas(a)redhat.com> wrote:
>
>
> ----- "Vladimir Benes" <vbenes(a)redhat.com> wrote:
> > Hi all,
> > we discussed yesterday the possibility of returning take button
into beaker. Maybe it can help something but as for now I use it like this.
> >
> > find myself any machine.. select the distro I want to provision and
then click schedule provision (in the certain system page). This creates
me a job and send me notification about it. So it works as well as take
and even better in a way of scheduling currently occupied system as you
don't have to check if it's already free.
> > The job is not cancelled when rebooting so you can make whatever
you want with the box.
> >
> > I don't see any reason for changing the code to get the Take button
back, any ideas what went wrong so the need to have Take button came out?
>
> There is one, probably not sufficient enough since it's not too
common case. Manual provisioning. If I want to install machine manually
through vnc, need to perform it when the job starts. With oldschool
provision I would reserve the machine and then reprovision it. This is
now impossible. I reserve the machine, enqueue job with vnc install and
return2beaker with hope no other jobs are in queue.
>
> I know I could use my own kickstart, but it seems to me it would cost
me much more time. Actually, I don't need Take button, but ability to
reprovision the machine.
Just an idea, what about "Schedule provision" for machines where you're
not current user and oldschool "Provision" if you're current user?
This would meet needs Jan and I were describing and the infinite time
reservation would remain deprecated, as was intention of Take button
removal. Would this be hard to implement? Reprovisioning the machine and
starting new reservesys with the time remaining...
>
> >
> > cheers,
> > Vladimir
> >
>
> --
> Petr Benas
> Kernel QA Brno
> RedHat Czech, s.r.o
> pbenas on #brno, #kernel-qe, #qa
>
--
Petr Benas
Kernel QA Brno
RedHat Czech, s.r.o
pbenas on #brno, #kernel-qe, #qa
13 years, 6 months
priorities when specific machine needed?
by Vladimir Benes
Hi all,
I am facing a weird thing. I need Dell PE 1950 but almost all jobs are
scheduled using those machines so it's really hard to get those.
I had one of them provisioned but after creating a new job (I needed 5U3
instead of 5U5) and returning the machine into the spool I was quite
surprised as scheduler immediately used that machine on another
no-machine-specified job w/o checking if someone in the queue needed the
specific machine.
This behaviour is quite dangerous as it can block people from getting
right machine. We had to (thanks mganisin) cancel the job two times to
get machine to me again.
bug created:
https://bugzilla.redhat.com/show_bug.cgi?id=635611
thanks,
Vld
13 years, 6 months
Re: [Beaker-devel] ask about some questions about beaker
by Raymond Mancy
Hi Junyou,
It doesn't have anything to do with this does it ?
https://fedorahosted.org/cobbler/ticket/503
Cheers
----- "钧友 杨" <smile.yangjunyou(a)yahoo.com.cn> wrote:
> From: "钧友 杨" <smile.yangjunyou(a)yahoo.com.cn>
> To: beaker-devel(a)lists.fedorahosted.org
> Sent: Friday, August 20, 2010 5:33:39 PM GMT +10:00 Brisbane
> Subject: [Beaker-devel] ask about some questions about beaker
>
>
>
Hi ,
I want to setup lab controller as Beaker-0.5-Administration_Guide-en-US.pdf, but when it comes to the step :
"$ cobbler import --path=/net/${NFSSERVER}/${NFSPATH} \ -- name=$DISTRONAME \ --availableas=
> nfs://${NFSSERVER}:/${NFSPATH}"
It just fails and shows that :
[root@localhost ~]# cobbler import --path=/root/ubuntu-10.04-rc-alternate-amd64 --name=ubuntu --arch=x86_64
> task started: 2010-08-20_152939_import
> task started (id=Media import, time=Fri Aug 20 15:29:39 2010)
> running: rsync -a '/root/ubuntu-10.04-rc-alternate-amd64/' /var/www/cobbler/ks_ mirror/ubuntu-x86_64 --exclude-from=/etc/cobbler/rsync.exclude --progress
> sending incremental file list
> rsync: change_dir "/root/ubuntu-10.04-rc-alternate-amd64" failed: Permission denied (13)
> rsync: ERROR: cannot stat destination "/var/www/cobbler/ks_mirror/ubuntu-x86_64": Permission denied (13)
> rsync error: errors selecting input/output files, dirs (code 3) at main.c(554) [receiver=3.0.5]
> rsync: connection unexpectedly closed (9 bytes received so far) [sender]
> rsync error: error in rsync protocol data stream (code 12) at io.c(600) [sender=3.0.5]
> returned: 12
> Exception occured: <class 'cobbler.cexceptions.CX'>
> Exception value: 'Command failed'
> Exception Info:
> File "/usr/lib/python2.6/site-packages/cobbler/utils.py", line 122, in die
> raise CX(msg)
Exception occured: <class 'cobbler.cexceptions.CX'>
> Exception value: 'Command failed'
> Exception Info:
> File "/usr/lib/python2.6/site-packages/cobbler/remote.py", line 95, in run
> rc = self._run(self)
> File "/usr/lib/python2.6/site-packages/cobbler/remote.py", line 224, in runner
> self.logger
> File "/usr/lib/python2.6/site-packages/cobbler/api.py", line 657, in import_tree
> return importer.run()
> File "/usr/lib/python2.6/site-packages/cobbler/action_import.py", line 181, in run
> self.run_this(rsync_cmd, (spacer, self.mirror, self.settings.webdir, self.mirror_name))
> File "/usr/lib/python2.6/site-packages/cobbler/action_import.py", line 262, in run_this
> utils.die(self.logger,"Command failed")
> File "/usr/lib/python2.6/site-packages/cobbler/utils.py", line 130, in die
> raise CX(msg)
!!! TASK FAILED !!!
> You have new mail in /var/spool/mail/root.
It's weird that it shows that I don't have the permission of /root/ubuntu-10.04-rc-alternate-amd64,but when I used "ls -l /root/",it shows that:
[root@localhost ~]# ls -l /root/
> total 80
> -rw-------. 1 root root 1656 2010-08-20 23:27 anaconda-ks.cfg
> -rw-r--r--. 1 root root 58748 2010-08-20 23:27 install.log
> -rw-r--r--. 1 root root 4843 2010-08-20 23:27 install.log.syslog
> drwxr-xr-x. 10 root root 4096 2010-08-20 15:11 ubuntu-10.04-rc-alternate-amd64
>
Thanks,
Junyou
>
> _______________________________________________ Beaker-devel mailing list Beaker-devel(a)lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/beaker-devel
13 years, 6 months
[PATCH] - Review for bz629272
by Bill Peck
Bugzilla 629272 explains that a loaned system can still be grabbed by
other users even though the loanee should have exclusive access while
the loan is in effect. I thought I had fixed this before but obviously
didn't. Looking at my previous code I removed the system from the list
of possible systems but were too late for that now, we've already made
our decision about which system to use. So the proper fix is to make
sure we don't execute the rest of the "I have a system available" code.
Simply added an elif.
commit 7f33ab29e7eac9ff60c1dd919a70844527be92ea
Author: Bill Peck <bpeck(a)redhat.com>
Date: Wed Sep 1 11:15:47 2010 -0400
bz629272 - fix logic to prevent us from using systems that we no
longer have access to.
diff --git a/Server/bkr/server/tools/beakerd.py
b/Server/bkr/server/tools/beakerd.py
index b3ce8a3..2e4e1c2 100755
--- a/Server/bkr/server/tools/beakerd.py
+++ b/Server/bkr/server/tools/beakerd.py
@@ -270,21 +270,21 @@ def queued_recipes(*args):
log.debug("System : %s is available for Recipe %s" %
(system, recipe.id))
# Check to see if user still has proper permissions to
use system
# Remember the mapping of available systems could have
happend hours or even
# days ago and groups or loans could have been put in
place since.
if not System.free(user).filter(System.fqdn ==
system).first():
log.debug("System : %s recipe: %s no longer has
access. removing" % (system,
recipe.id)
recipe.systems.remove(system)
# Atomic operation to put recipe in Scheduled state
- if session.connection(Recipe).execute(recipe_table.update(
+ elif
session.connection(Recipe).execute(recipe_table.update(
and_(recipe_table.c.id==recipe.id,
recipe_table.c.status_id==TaskStatus.by_name(u'Queued').id)),
status_id=TaskStatus.by_name(u'Scheduled').id).rowcount == 1:
recipe.createRepo()
13 years, 6 months