Cheers,
>>
>> Fritz
>>
>>> Am 04.08.2016 um 07:01 schrieb Simon Matthews :
>>>
>>> Has anyone constructed a cluster using Kubernetes, Docker and
>>> Gridengine (SGE or newer variants)? Does this even make sense? Is
>>> there a better set of tool
Just to add 2 points to what Chris wrote: The amount of cores is less
important. What's important is enough memory, fast access to storage where
spooling goes and, on a really busy system, the performance of the cores.
Depending on the exact workload profiles and policy set-ups one can become
more
>
>
>
> ___
> users mailing list
> users@gridengine.org
> https://gridengine.org/mailman/listinfo/users
Fritz Ferstl | CTO and Business Development, EMEA
Univa Corporation <http://www.univa.com/> | Th
Hi Sylvain,
Univa has create a native integration into Univa Grid Engine which we
call Container Edition. It is in Beta right now and will release by end
of this month.
You can do something like a 'docker run' inside of a job script even
without any integration, of course, but you'll loose a
>> On Sun, Apr 24, 2016 at 12:36 AM, Fritz Ferstl wrote:
>> Hi Mark,
>>
>> I'll respond via separate e-mail to you because I do not want to misuse this
>> mailing list for too much commercial messaging. On behalf of other readers
>> who may have sim
Hi Mark,
I'll respond via separate e-mail to you because I do not want to misuse
this mailing list for too much commercial messaging. On behalf of other
readers who may have similar requirements I would, however, like to
provide a brief overview of Univa's product line-up and how it relates
t
One other thing that could be done is create the starter method for the
"old" queue that resubmits the job while modifying any occurrence of -q
or -l q=, etc.
That is kind of an alias then but also not too different from a wrapper
script.
Cheers,
Fritz
Tina Friedrich schrieb:
On 10/03/16
One could also wrap the shepherd if what you wanted to do is check for
the working directory and potentially create or mount it if it isn't
there yet before exec'ing the real shepherd. All so called methods
(prolog, starter, pe-*, etc) are run by the shepherd. So by wrapping it
you can precede
There actually is a deadline policy and given you wrote the jobs
approach a deadline it might be worth considering it. The job needs to
be submitted as deadline job, however, and (out of the top of my head)
the following conditions need to be met for it to work:
* The user needs to be a membe
Hi Atul,
I will respond to you via personal e-mail regarding your Univa Grid
Engine evaluation.
If you have general questions around how checkpointing works in
combination with Grid Engine in general then we can also discuss this on
the community alias because others might benefit too or cou
influence of the functional policy relative to others
* Now submit your jobs with '-js '
Maybe this helps to achieve what you guys are looking for.
Fritz
Fritz Ferstl schrieb:
Oh ... and you can also mix the functional with the fairshare policy ...
Fritz
Fritz Ferstl schrieb:
Oh ... and you can also mix the functional with the fairshare policy ...
Fritz
Fritz Ferstl schrieb:
I'm always interested ;-) - hence I'm still involved after more than
20 years ...
As I wrote, I'd look into the functional policy.
Or if you are using fairshare, as Chris w
lar to what you want.
Cheers,
Fritz
Mark Dixon schrieb:
On Thu, 30 Apr 2015, Fritz Ferstl wrote:
Nah, the weight_priority won't help. It just determines how much
influence the -p has vs things like job wait time or urgency. If you
have none of those then all being equal it would have th
Nah, the weight_priority won't help. It just determines how much
influence the -p has vs things like job wait time or urgency. If you
have none of those then all being equal it would have the same effect as
if you left it untouched. And if you have influence from waiting time or
urgency or othe
Hi Chris,
if memory serves correctly then the so called POSIX priority is one
range applied to all jobs. So you would be penalizing your jobs.
If you need this more often then you could either ensure that jobs
always are submitted with a negative priority (say -100) and then
regular users ha
It is certainly an intended feature, William. Always was since load
sensor were introduced in the late 90s.
The thought behind it was that you might have central system management
services which maintain host level information. You can then put the
load sensor on the system management server i
Hi Arne,
quite simply SoGE has been using Univa's 8.0.0 code we have made
available publicly. After that there was no relationship. So while there
were similar release numbers for some time (until we have released Univa
Grid Engine 8.2.0) the contents are different.
The various Univa Grid En
Hi,
for those of you who are tracking how Univa Grid Engine is evolving it
should be interesting to note that we have released 8.2.0. Some key
features are:
* Native Microsoft Windows support (not requiring compatibility layers
like SFU or cygwin and working on recent Windows versions)
*
a result I would like to have a more direct python interface to control
> grid engine.
>
>
>
>
>> On 25 June 2014 13:49, Fritz Ferstl wrote:
>>
>>
>>> Am 25.06.2014 um 14:31 schrieb Ashley Retallack :
>>>
>>
>>> I take
r is somehow embedded? Does
that matter?
Cheers,
Fritz
>
> If we aren't updating GridEngine it really shouldn't matter though surly?
>
>
>
>> On 25 June 2014 12:25, Fritz Ferstl wrote:
>> Reuti is right. The GDI does change ... If the SW internals itself do
Reuti is right. The GDI does change ... If the SW internals itself do undergo
bigger changsd (which is the case on our end). One other reason for why we're
wrapping a web services interface around it.
Cheers,
Fritz
> Am 25.06.2014 um 12:53 schrieb Ashley Retallack :
>
> I know, but we aren't
work. The web services interface isn't less work but
we're doing it for multiple reasons and then wrapping Python around that is
straight forward.
Let me know whether you're interested.
Cheers,
Fritz
--
Fritz Ferstl - CTO, Univa Corporation
ffer...@univa.com
> Am 25.0
uot;s%^.*@\([^<]*\)<.*%\1%"
In "qhost -j" (without the -xml) the length of the hostname field is
limited to 21 characters. So even that might be enough.
Cheers,
Fritz
Fritz Ferstl schrieb:
Tim,
the XML output returned with -xml (e.g. in addition to 'qstat -u ...&#
Tim,
the XML output returned with -xml (e.g. in addition to 'qstat -u ...')
will contain the un-truncated data. Should be easy to parse it out.
But indeed ... it should be much easier to get this info.
Cheers,
Fritz
Tim Landscheidt schrieb:
Ian Kaufman wrote:
When I run qstat -u \*, eac
should
be OK that I post the link to our trainings despite that they are for a
cost:
http://www.univa.com/resources/upcoming-training.php
Hope to see some of you there.
Best regards,
Fritz
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center
/releases.php.
Our free trial version should soon be updated to 8.1.7 in case you want
to test drive CGROUPS support or some of the other features.
Cheers,
Fritz
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center Optimization CompanyE-Mail:
ffer
Fritz Ferstl
7. November 2013
18:26
Hhhhmm ...you could use a
job context variable (see the qsub/qalter man page on job context vars
and the options to set them) for instance as a counter to ascertain in a
prolog that a particular task really is the last from an
erver or even prolog... (if possible).Anyone
with something more elegant? TIA,Arnau___users
mailing listusers@gridengine.orghttps://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Bu
ship of all copyright and
trademarks, _including_ the code under the SISSL, that Univa will be
fighting to shutdown the open source versions of Grid Engine (open
gridscheduler and Son of Grid Engine)?
John
Kloss II.
_______users mailing
listusers@g
early to ask this, but will more of the IP be moved
to SISSL or other open license in the near term or over time?Thanks,Bill
Fritz Ferstl
22. Oktober 2013
15:51
Yes, Mark. This is in
essence what I was intending to respond. The SISSL is a recognized and
liberal open s
ISSL, that Univa will be
fighting to shutdown the open source versions of Grid Engine (open
gridscheduler and Son of Grid Engine)?
John
Kloss II.
___users mailing
listusers@gridengine.orghttps://gridengine.org/mailman/listinfo/users
Fr
technology and also for
the Grid Engine team at Univa.
If you wish to read more about this please see the press release here:
http://www.univa.com/about/news/press_2013/10222013.php
Best regards,
Fritz
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data
oab due to poor performance.
Anyone else get this? I'm wondering if an over zealous sales rep is
scraping our mailing list archives or something ...
-dag
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/use
orums are free of charge and
the trainings are for a fee. Both can be attended separately.
See here for more information:
http://www.univa.com/resources/grid-engine-user-forum.php
Hope to see you there!
Cheers,
Fritz
--
Fritz Ferstl, CTO, Univa Corporation
m: ffer...@univa.comp: +49.170.819
_
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center Optimization CompanyE-Mail:
ffer...@univa.com | Phone: +49
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center Optimization CompanyE-Mail:
ffer...@univa.com | Phone: +49.9471.200.195 | Mobile:
+49.170.819.7390
___
w can this be achieved?
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center Optimization CompanyE-Mail:
ffer...@univa.com | Phone: +49.9471.200.195 | Mo
istinfo/users
_______
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center Optimization CompanyE-Mail:
ffer...@univa.com |
customer, although the particular cluster I was looking
> at hasn't yet been migrated to UGE. I'm not sure if we have a
> licenses/support for UniSight though.
>
> --
> Brian McNally
>
> On 05/09/2013 12:32 AM, Fritz Ferstl wrote:
>> Jesse is right that it is &qu
which position
their jobs are in the pending list (usually it's switched off for
performance reasons).
-- Reuti
Thanks,
--
Brian McNally
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
_____
nstitute
3601 Spruce St.
Room 214
Philadelphia PA 19104
Tel: 215-495-6914
Sending me a large file? Use my secure dropbox:
https://cscb-filetransfer.wistar.upenn.edu/dropbox/btay...@wistar.org
--
Fritz Ferstl | CTO and Business Development, EMEAUniva
Corporation | The Data Center Optimization
Not sure I have seen this guy from UCI. Runs SoGE.
Fritz
Sent from my iPhone
Am 31.03.2013 um 05:41 schrieb Adam Brenner :
> Howdy GE Users,
>
> We are running GE 8.1.2. I have setup a consumable resource on our
> cluster that will monitor available local storage on each of our
> execution no
Hi Gaya,
I see you are from Edinburgh and Univ of Edinburgh happens to be a Univa Grid
Engine customer. If you're part of that cluster and you are in fact using Univa
Grid Engine then feel free to get your questions answered by our support. We
can take it off-line if you've questions around tha
project by itself so that "bio" project
> does not interfere with "math" project?In other words, a FairShare setup
> only for "bio" users and a different FairShare setup for "math" users?
>
> Joseph
>
>
>
>
>
> ___
Hi Chris,
no, you cannot currently extend h_rt.
We're going to change that in an upcoming release but this doesn't help you
with your job ...
Fritz
Sent from my iPhone
Am 15.11.2012 um 09:20 schrieb Chris Dagdigian :
>
> Quick question ...
>
> I've got a job with a user running in a queue
ust.
Hope to see you soon and, if you attend, then consider making some room for a
quick 1.5 hr trip to Munich to the Oktoberfest ;-)
Cheers,
Fritz
Fritz Ferstl | CTO and Business Development, EMEA
Univa Corporation | The Data Center Optimization Company
E-Mail: ffer...@univa.com | Phone: +49.94
like to urge you to register!
You will notice that there also is the option to present interesting use cases,
integrations or similar topics. Please send us your proposals as detailed in
the above web page.
Looking forward to seeing you soon (again)!
Fritz
Fritz Ferstl | CTO and Business Develo
imple fact it is
> no more in qstat mean everything was ok?
>
> Thanks in advance
>
> regards
>
> Lionel
> ___
> users mailing list
> users@gridengine.org
> https://gridengine.org/mailman/listinfo/users
Fritz Ferstl | CTO an
Not sure that's the problem, William. The masterq-switch is pretty broken in
the most versions. Does it work as desired if you specify either of the two
variants (without trying to override)? If not then it might simply be the
broken masterq switch and you were misled to think it's the command-l
Am 25.05.2012 um 16:01 schrieb William Hay:
>
>
> On 25 May 2012 14:45, Fritz Ferstl wrote:
> Hhhhmmm ... would need to look into the code whether it even would complain
> about that for a queued job. It's clear for a running job and if that job was
> running before a
hrieb William Hay:
> On 25 May 2012 14:00, Fritz Ferstl wrote:
>> Hi William,
>>
>> remove it from the queue's pe_list. Then wait for the job(s) to finish and
>> you can delete it.
>
> I deleted the only queue that referenced it. There were no jobs in
>
Hi William,
remove it from the queue's pe_list. Then wait for the job(s) to finish and you
can delete it.
If you really wanted to modify the PE then you can do the same and setup the
new PE with modifications while the old one is still there but no longer
referenced by the queue.
Cheers,
Fri
rly braces".
> I was thinking of something like:
>
> $ eval {x,y,z}=1
Yep. We did put out the draft of the spec for public comments at the time and
were taking inputs into account before implementing it. So that's how it ended
up becoming what it is today ...
Cheers,
Fritz
s here recently, and
> there seems to be plenty that needs a response if I can find time.
> Anyhow, now for something more constructive.]
> _______
> users mailing list
> users@gridengine.org
> https://gridengine.org/mailman/listinfo/users
Fritz Ferstl | CTO and Business
---
>> ___
>> users mailing list
>> users@gridengine.org
>> https://gridengine.org/mailman/listinfo/users
>
>
> ___
> use
estion is which version?
We had Fritz Ferstl in yesterday to explain Univa's commercial
offering and what they offer over and
above the common core. Fritz also explained that contrary to our
experience Grid engine is really
buggy, unstable and hard to debug so we really need a support contra
in the www
directory of
the repo if I recall correctly.
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business
y be that what they are asking is not directly possible, in
which case you can propose reasonable alternatives.
--
Fritz Ferstl | CTO and Business
Development, EMEA
Univa
alues h_vmem=($mem_total-0.5G)
Is something like this possible?
regards, Andreas
--
Fritz Ferstl | CTO and Business
Development, EMEA
ebian and Fedora having to patch it, at
least, but I don't understand all their current patches.
--
Fritz Ferstl | CTO and Business Development, EMEA
Univa Corporation <http://www.univa.com/> | The Data Center Optimization
Company
E-Mail: ffer...@univa.com | Phone: +49.9471.200.195 | Mobile:
+4
ractive'
and 'batch' work suitably as queues?
I've picked up a lot of good information in the past couple of weeks.
Now I need to get some time to work with things and try some more
experiments.
Thanks,
Stuart
--
rg/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business
Development, EMEA
Univa
Corporation | The Data Center Optimization Company
E-Mail:
ffer...@univa.com | Phone: +49.9471.200.195 | Mobile:
Am 02.05.11 20:31, schrieb Dave Love:
Fritz Ferstl writes:
If you look at the approved power-down cycles which the vendors
publish for state-of-the-art HW then that may be scary. It usually is
in the order of 2000 cycles.
Indeed, thanks
__
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business
Development, EMEA
Univa
Corporation | The Data Center Optimi
--
Fritz Ferstl | CTO and Business
Development, EMEA
Univa
Corporation | The Data Center Optimization Company
E-Mail:
ffer...@univa.com | Phone: +49.9471.200.195 | Mobile
gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business
Development, EMEA
Univa
Corporation | The Data Center Optimization Company
E-Mail:
not for
holding long term detailed information.
If you can get it into a host value via a load sensor, presumably it
could go in the dbwriter database.
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
Yep. That's why I mentioned high availability.
Cheers,
Fritz
On Mon, 11 Apr 2011 at 7:34pm, Fritz Ferstl wrote
I guess that "over NFS" makes the difference here.
Spooling over NFS is not the best thing to do perfor
BDB. Are there other tunables I should be
looking at?
--
Fritz Ferstl | CTO and Business
Development, EMEA
Univa
Corporation | The Data Center Optimization Company
E-Mail:
ff
another thing I did was -b y vs n, and on Linux it
also did not make much difference with the local disk.
Rayson
On Mon, Apr 11, 2011 at 12:28 PM,
Fritz Ferstl <ffer
ridengine.org
https://gridengine.org/mailman/listinfo/users
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business
Development, EMEA
U
ome.
-dag
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl | CTO and Business
Development, EMEA
Uni
Am 23.03.11 20:05, schrieb Dave Love:
Chiara Latini writes:
I'm planning to get some of the informations I need by connecting to the
ARCo DB from my Python code, and to invoke the GE commands once I got
the data and I've taken the proper decisions.
Right. That's what I'd suggest, without
_
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl -- CTO and Business Development, EMEA
Univa -- The Data Center Optimization Company
E-Mail: ffer...@univa.com
Web: http://www.univa.com
Phone: +49.9471.20
cally searchable on
sunsource either.
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
___
users mailing list
users@gridengine.org
https://gridengine.org/mail
Hi Rayson,
Am 16.03.11 18:44, Rayson Ho wrote:
On Wed, Mar 16, 2011 at 12:28 PM, Fritz Ferstl wrote:
Sure we can merge things but is that the spirit of an open source project?
That one party only touches the code?
I don't think this statement made any reference to the existence of
Am 16.03.11 15:47, Dave Love wrote:
Fritz Ferstl writes:
We'll welcome improvements and are definitely willing to assist with
advice but we can't dedicate more to effort to this right now.
I wasn't meaning you should do it. Could we perhaps see the existing
Univa configurat
dengine.org https://gridengine.org/mailman/listinfo/users
--
Fritz Ferstl -- CTO and Business Development, EMEA
Univa -- The Data Center Optimization Company
E-Mail: ffer...@univa.com
Web: http://www.univa.com
Phone: +49.947
Am 16.03.11 16:10, Dave Love wrote:
Fritz Ferstl writes:
It should actually be quite easy. In a first implementation you'll
probably want to introduce a qmaster_param "black_hole_exit_rate" and
then keep a statistic of the exit frequency rate for each exec host near
the cod
Am 12.03.11 00:49, schrieb Dave Love:
Rayson Ho writes:
Fritz,
So is there an internal branch of Grid Engine source tree inside the
Univa firewall?? And if that's the case, how long does it take for
changes in the internal branch to reach the outside branch that
everyone can checkout??
You
Am 12.03.11 00:39, schrieb Dave Love:
Fritz Ferstl writes:
Binaries based on the open source need to become a community effort.
What really needs to be done is to make it rather less challenging to
build, and then the issue should largely go away.
We're fully aware of that a
Am 12.03.11 01:20, schrieb Dave Love:
Fritz Ferstl writes:
It's something that definitely needs to be added to Grid Engine (as a
built-in feature) and we'll do that sooner rather than later but it
would be a nice little project for community members who have not done
code contribu
I'll wait a month come back to see where the user community is.If
Univa decides not to provide a courtesy binary for MacOS, I'll pitch
in.
MacOS is currently not on our (Univa's) platform support matrix. We have
not locked down the matrix, though. If we receive customer requests then
we c
It's something that definitely needs to be added to Grid Engine (as a
built-in feature) and we'll do that sooner rather than later but it
would be a nice little project for community members who have not done
code contributions thus far (hint! ;-)).
Rayson is correct that you'll need a heurist
ing else and that is the reality.
Cheers,
Fritz
Am 17.02.11 12:25, Reuti wrote:
Am 17.02.2011 um 09:07 schrieb Fritz Ferstl:
My 2 cents here (and I'm aware they will not help Eric ... and apologies in
advance for the rant, it's a long-term heartfelt topic ...):
A DRM benchmark
z
Am 17.02.11 10:52, schrieb Andreas Haupt:
Hi Fritz,
On Mon, 2011-02-14 at 20:24 +0100, Fritz Ferstl wrote:
Univa is also conducting a series of free webinars informing interested
parties about our Grid Engine roadmap. Note that attendance is limited
and the webinars are requiring registration.
My 2 cents here (and I'm aware they will not help Eric ... and apologies
in advance for the rant, it's a long-term heartfelt topic ...):
A DRM benchmark would be nice to have. Benchmarking in general is an
almost vain attempt. You have to be very prescriptive of the boundary
conditions to ach
Hi,
please don't take this as a shameless advertisement. I wouldn't post
this if the offers listed below hadn't been tailored to this audience
and if they wouldn't include free access options. So I trust the offers
described here are interesting for at least some of you:
http://www.univa.com
88 matches
Mail list logo
