you need to keep the indentation!
(For the full story, read:
https://github.com/gc3-uzh-ch/elasticluster/issues/304 )
Ciao,
R
--
Riccardo Murri, Schwerzenbacherstrasse 2, CH-8606 Nänikon, Switzerland
--
You received this message because you are subscribed to the Google Groups
"elasticlus
Dear Ana:
the `image_userdata` parameter is only used when *starting new VMs*.
Did you destroy (stop) the cluster and make a new one afresh?
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop
d-Upgrade "0";
__EOF__
* save the running VM as a snapshot
* use that snapshot in the ElastiCluster config, in stead of the official image
Hope this helps,
R
--
Riccardo Murri, Schwerzenbacherstrasse 2, CH-8606 Nänikon, Switzerland
--
You received this message because you ar
Dear Ana:
> Yes, I did stop the cluster and started new one.
Hmmm... then going back to the error message you posted last, I see two
distinct issues:
1. SSH keys mismatch (not a fatal error)::
fatal: [compute001]: FAILED! => {... "module_stderr": "
... WARNING: REMOTE HOST
size -r` removes the nodes immediately
starting with the highest-numbered ones, regardless of whether they are
running any jobs.)
Ciao,
R
--
Riccardo Murri, Schwerzenbacherstrasse 2, CH-8606 Nänikon, Switzerland
--
You received this message because you are subscribed to the Google Groups
&quo
lation should have the
EC2 compatibility layer installed; you should see the `ec2-api-metadata`
service running and listening to port 8788 or 8789 if it does.)
Are you able to (1) start a VM with the same keypair and VM image used by
ElastiCluster and (2) ssh into it from the command-line?
Ci
VM images, users with access to the same UNIX
account used by Ansible, etc) then you can download the release
candidate Ansible code from: http://releases.ansible.com/ansible/
Ciao,
R
--
Riccardo Murri, Schwerzenbacherstrasse 2, CH-8606 Nänikon, Switzerland
--
You received this message beca
Ok, *now* the Ansible package with fixed code has landed on PyPI,
so ElastiCluster follows suit and requires Ansible>=2.2.1.0
Sorry for the confusion!
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and
Dear ElastiCluster users,
On 5 September 2016 at 14:21, Riccardo Murri <riccardo.mu...@gmail.com>
wrote:
>
> I'm still aiming at supporting Py 2.6 until (and including) the 1.4
> release of ElastiCluster (scheduled for Nov. 2016), but I would like to
> drop it beginning of 20
iCluster version 1.4,
planned for end of November.
Thanks,
Riccardo
--
Riccardo Murri, Schwerzenbacherstrasse 2, CH-8606 Nänikon, Switzerland
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop receivi
Hello,
since a couple of days, all compute clusters installed with
ElastiCluster run a "hpc-common" Ansible role, the purpose of which is
to install software and configuration that people would normally
expect on a HPC-oriented cluster.
At the moment, the role is quite bare-bones and only
Hi Pablo,
> thanks for your help. The error appeared in the "Gathering Facts" task in
> ansible.
What version of Ansible are you using?
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscrib
nor Ansible are to blame -- ElastiCluster's code is
:-(
Could you please report this issue on GitHub?
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
T
ter.readthedocs.io/en/latest/playbooks.html
Good idea, I'll try to split that page into "fundamental" playbooks and
"add-on" ones.
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed t
ing anyone?
Thanks for any comments!
[1]:
https://specs.openstack.org/openstack/openstack-specs/specs/clouds-yaml-support.html
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed to the Google Groups
&quo
you create your own VM
image/snapshot, which includes the startup script and calls it from
`/etc/rc.local` or a similar boot-level script.
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed to the Google
'module' object has no attribute 'get_configurator'
>
> I'm guessing `get_configurator()` is from an older version
> elasticluster?
Yes, the configuration parser was refactored and that function removed
in commit 94859c4 on 2016-10-12.
I'll check what the `cluster_monitor.sh` code is d
email me or the list! (e.g.: "using ElastiCluster
for creating short-lived Hadoop clusters (size 8 to 20 nodes) for
running one-off analytics tasks", or "using ElastiCluster to create
clusters for teaching", etc.)
Thanks,
R
[*]: https://indico.cern.ch/event/595396/contributions/25
For all those interested: this issue is being discussed on GitHub:
https://github.com/gc3-uzh-ch/elasticluster/issues/473
Please follow-up there.
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop
Hi Pablo,
It is probably an issue with the code: I renamed an attribute
'nova_api_version' into 'compute_api_version' but possibly missed an
occurrence. Or Maybe it's the saved state of the cluster that should
converted.
Anyway, going back in the sources to a couple days ago should fix it.
I'll
Hi all,
many thanks for the feedback! So it looks like the consensus is for
making more features opt-in.
Good to know :-) I'll work towards that (as usual, it may take a while).
If you already have any patches for this, please feel free to submit a
PR on GitHub!
Ciao,
R
--
Riccardo Murri
m/gc3-uzh-ch/elasticluster/issues/480
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop receiving emails
--
Riccardo Murri
http://www.s3it.uzh.ch/about/team/#Riccardo.Murri
S3IT: Services and Support for Science IT
University of Zurich
Winterthurerstrasse 190, CH-8057 Zürich (Switzerland)
Tel: +41 44 635 4208
Fax: +41 44 635 6888
--
You received this message because you are subscribed to the Google
Engine users mailing list is a good place to ask questions on
Grid Engine: https://gridengine.org/mailman/listinfo/users
Hope this helps,
Riccardo
--
Riccardo Murri, Schwerzenbacherstrasse 2, CH-8606 Nänikon, Switzerland
--
You received this message because you are subscribed to the Google Grou
he GE variant provided by the base OS.
Perhaps something has changed upstream (e.g., location of `act_qmaster`)
which broke the playbooks.
I'm on vacation now, but I will back to my office on Monday and can have
a look if you have not solved the issue by then.
Ciao,
R
--
Riccardo Murri, Schwerzenba
Hello John-Paul,
is this still an issue? I've copied it to ElastiCluster's GitHub
issue tracking page:
https://github.com/gc3-uzh-ch/elasticluster/issues/452
Sorry for the late reply!
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
Hello Alexander,
welcome to ElastiCluster :-)
Yes, ElastiCluster does automatically provision and configure a
cluster with SLURM (or other systems, see list at [1]): after you've
done the initial configuration, you should be able to get a
fully-functional SLURM cluster on EC2/GCE/OpenStack/etc.
2017-09-21 22:36 GMT+02:00 Alex :
> so as long as i follow the ElastiCluster
> setup instructions i won't have to do anything else? (ie i won't have to
> actually install slurm separately and configure it)
Yes, exactly.
Cheers,
Riccardo
--
You received this
Hello,
Traceback (most recent call last):
File ".../elasticluster/cluster.py", line 499, in _start_node
node.start()
File ".../elasticluster/cluster.py", line 1141, in start
**self.extra)
File ".../elasticluster/providers/ec2_boto.py", line 197, in
start_instance
connection =
Hello,
supporting Python 2.6 is becoming next to impossible; even Paramiko and
Pip have dropped support for it. Is there anyone using ElastiCluster
that still needs Python 2.6 compatibility? (and cannot use other
options, like a non-system Python 2.7 or Docker?)
Ciao,
R
--
Riccardo Murri
Hello Samy,
any success in using the "official" OS images with ElastiCluster?
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
Hello,
> I tried again and I'm getting different errors. Please see the attached log.
>
> Do you recommend another OS version? I'm almost giving up.
The AMI you're using seems to lack the package `lua-devel`,
which however *is* part of a standard CentOS7 (see e.g.
Hello,
> So yes, I would suggest you switch to a "standard" AMI;
Important: the AMI ID is recorded in the cluster upon creation -- so
changing the AMI in the configuration file is not enough. You need to
destroy a cluster (`elasticluster stop -y myaws`) and then re-create
it.
Ciao,
R
--
You
Hi Champak,
You likely forgot to source the `openrc` file before running Elasticluster.
Ciao,
R
Il mar 8 mag 2018, 19:35 Champak Reddy ha scritto:
> Hi Riccardo,
>
> I was trying to update my cluster with the "setup" command and I get the
> errors below.
> *$
Hi Orxan, all,
> Elasticluster spent nearly two hours for configuration of a cluster with 37
> nodes.
Yes, this is definitely a pain point with ElastiCluster/Ansible ATM.
I'll try to
summarize the issue and give some suggestions here.
My rule of thumb for time it takes to set up a basic SLURM
Hello,
it looks like the system gets into trouble when installing package `moreutils`::
failed: [compute004] (item=moreutils) => {"changed": true,
"failed": true, "item": "moreutils", "msg": "Error: Package:
moreutils-0.49-2.el7.x86_64 (epel)\n Requires:
perl(IPC::Run)\n", "rc": 1,
As far as I can see, OpenStack provisions "raw" devices as
volumes.
The Ansible "filesystem" module may help you with this:
http://docs.ansible.com/ansible/latest/filesystem_module.html
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
ou please run the following commands on the compute nodes to
check::
rpcinfo -p localhost
rpcinfo -p frontend001
showmount -e frontend001
Thanks,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed
Dear fellow EasltiClusters,
Would you have time to test the new [elasticluster.sh][1]
installation-free script?
Just download it anywhere on your path, and use it like the regular
ElastiCluster::
wget -O elasticluster.sh
Ah, no, wait! You should *not* run `elasticluster.sh` through `sudo`!
Otherwise you'll be running as root in your home directory, which
screws all permissions up...
Can you please run the following commands?
# fix permissions
chown -R $USER $HOME/.ansible $HOME/.ssh $HOME/.elasticluster
Dear Orxan,
there seems to be an error with the Docker image; according to this
log line, the Ansible configuration system did not run at all:
ansible.errors.AnsibleError: Unable to create local
directories(/home/.ansible/tmp): [Errno 13] Permission denied:
'/home/.ansible'
This is
2018-02-04 12:15 GMT+01:00 Orxan Shibliyev :
> The second command gave:
>
> orhan@orhan-MS-7850:~$ ./elasticluster.sh -vvv start slurm-on-gce
> docker: Got permission denied while trying to connect to the Docker daemon
> socket at unix:///var/run/docker.sock: Post
>
Hello Orhan,
> Sorry for taking your time with these fail messages but I got yet another
> one. Output is attached.
The issue is that two nodes were interrupted during `dpkg`'s configure
step (due to Ubuntu's obnoxious "upgrade on boot" setting), so they
refused to install any new software.
Dear all,
I have updated ElastiCluster's Ansible playbooks to use the new syntax
introduced in Ansible 2.4 and 2.5. This gets rid of the many
deprecation warnings that were littering ElastiCluster's output, but
also introduces some incompatibility with Ansible 2.3. So:
* Ansible 2.5 or 2.6 is
Hello!
I've just committed additions to the ElastiCluster playbooks to
install "R Studio Server" -- it can be combined with any cluster
software, or used stand-alone. See an example config to deploy it on
Google Cloud Engine:
# ... everything as usual here, plus:
global_var_upgrade_packages=no
Thanks!
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this grou
nstead of using the Docker image? (see:
http://elasticluster.readthedocs.io/en/latest/install.html#quickstart )
Ciao,
R
--
Riccardo Murri / Email: riccardo.mu...@gmail.com / Tel.: +41 77 458 98 32
--
You received this message because you are subscribed to the Google Groups
"elasticlu
Hello Shane,
I have seen the issue you filed on GitHub; I'm replying here but
either forum is equally good.
What is not clear to me is: did you install these 20 machines with
ElastiCluster? (I'd say
no, since you seem not to have an ".elasticluster/config" file...)
Or rather did you install all
Hello!
> First of all thank you very much for such a wonderful software.
Glad you like it! And thanks for getting back with questions -- this
is how a community is built and grows :-)
> 1. Is there any way already implemented for load balancing in Elasticluster?
I'm not sure what you mean by
Hello Orxan,
> it. I think this is related to google but after some research I am still
> clueless. Basically, I get following.
>
> Your browser has been opened to visit:
>
>
>
> If your browser is on a different machine then exit and re-run this
> application with the command-line parameter
Hello Orxan,
I cannot reproduce this error; with a freshly-started Ubuntu 16.04
cluster, I get::
ubuntu@frontend001:~$ cat test.sh
#! /bin/sh
echo hello
ubuntu@frontend001:~$ sbatch test.sh
Submitted batch job 2
ubuntu@frontend001:~$ cat slurm-2.out
hello
One caveat: right after building the
Hello Champak,
> I have successfully setup a multiuser cluster using Elasticluster with
> multiple nodes with differing architectures but I haven't been able to run
> more than a single job per node. p...]
> champost@frontend001:~/sbatch$ for i in `seq 20`; do sbatch -J test_$i -e
> test_$i.out
Hello Samy,
I haven't tried to build a cluster with the config you posted but my
first guess is that you cannot connect because of firewall rules in
EC2: i.e., your default security group does not allow SSH connections
from the host you're running ElastiCluster on.
(See, e.g., this comment for
Hello!
2018-04-15 17:21 GMT+02:00 'Ravi Arya' via elasticluster
:
> 1. Load balancing: You submit the number of jobs to the cluster and
> depending upon the number of jobs, cluster adds or removes the nodes. This
> is automatic and is accomplished by running load
Hi Orxan,
I tried starting a cluster with your config, but I still cannot
reproduce the error:
* the only differences are:
- I only started 1 compute node
- I used dthe public ebian-9-stretch-v20180401 as a base OS since I
cannot use the image-23 snapshot from your project
- I use my own
Probably a network glitch. Can you please try again and open an issue on
GitHub if the problem persists?
Ciao,
Riccardo
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop receiving emails from it, send an
Dear all,
I have just merged the new Azure provider into the "master" branch;
this means the Docker image will get working Azure support in a few
minutes.
The configuration information has changed; please see
Hi Champak,
> It worked like a charm. I am able to activate/deactivate the virtualenv and
> import elasticluster!
Excellent!
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop receiving emails
Hello all,
I was thiinking of adding a section to the ElastiCluster
documentation, to cover topics that are possibly interesting to users
and/ot needed in day-to-day usage of ElastiCluster but do not fit well
into the current sections installation/configuration/usage/playbooks.
Is there anything
Hi Champak,
> Is it possible to configure a cluster consisting of several compute nodes
> but with different flavours (e.g. made up of 8cpu-64ram-hpc and
> 4cpu-16ram-hpc and 1cpu-4ram-hpc flavors).
Yes, totally possible!
I've just added an example config here:
Hello,
> TASK [common : Upgrade all installed packages to latest version]
> fatal: [frontend001]: FAILED! => {"changed": true, "failed": true, "msg":
> "https://yum.repos.intel.com/2018/intel-daal-runtime-64bit-2018.2-199.x86_64.rpm:
> [Errno 14] curl#18 - \"transfer closed with 138280031 bytes
Hi all,
for the record, I'm summarizing here the outcome of the debugging
session which went on through private emails.
The issue stems from the user name on GCP vs user name on the VM:
Google uses the local part of the email address (e.g., in my case
`riccardo.murri`) but then the user name on
> But the cluster didn't get setup. I ran `elasticluster setup mycluster'
> various times and it fails. I'm using the same config file.
Then there is some other error before which is the root issue.
Can you please post the *entire* output of `elasticluster -v setup mycluster` ?
Ciao,
R
--
You
> Ans this is another error message when the script finishes:
Yes, this is just a consequence of the error that you hit before.
ElastiCluster is just summarizing the outcome of the setup process;
since there were fatal errors before, it warns you that the cluster has
not been fully configured.
Hmmm... you're right, installation of `moreutils` on CentOS 7 was
handled separately.
Should be fixed now -- can you pls update and check again?
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop
Dear all,
2019 starts with a long-awaited bug fix for Azure users:
ElastiCluster's "master" branch now has fixes for all the issues
Manuele reported!
Please test and let me know if it works (it does for me...)
Ciao & happy new year,
Riccardo
--
You received this message because you are
Hello Manuele, all,
> Thanks for the update. After a (long) round of tests I can confirm that
> ElastiCluster works with Azure very well now. I'm able to set up and manage
> my clusters!
Excellent! Thank you very much for taking the time to test!
Ciao,
R
--
You received this message
Hello Manuele,
I think the root cause of the problem is the exception thrown by the Azure
code -- the malformed YAML is just a consequence of that.
I'll take a look in the coming days; right now, my Azure account is
non-functional.
Thanks for reporting this issue!
Riccardo
--
You received
> Please disregard my previous post. I didn't even construct cluster but
> just one instance. Sorry for taking time.
>
All is well that ends well ;-)
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and
Hello Orhan,
> When I run a job by setting number of nodes to 1 and number of tasks to 1 as
> well, naturally, only compute001 runs the job. Then I run 8-node job and I
> see that output of first job which ran on compute001 are also available on
> other nodes. Does elasticluster copy files
Hello Orhan,
> So when a node produce a file, the file will be copied to all other nodes,
> right?
Almost. All files in the home directories actually reside on the
front-end node;
whenever you try to read or write to a file in the home directory,
this is transparently
routed to the front-end
Dear ElastiCluster users,
I'm pleased to announce that ElastiCluster now can install Hadoop
2.8.4 and Spark 2.2.1, through the Apache BigTop 1.3.0 (released
yesterday).
If you want to install the newer Hadoop/Spark, be sure to pull the
latest ElastiCluster from the "master" branch.
Ciao,
R
--
Hello!
ElastiCluster now has support for installing PBSPro [1], albeit only
on CentOS7 (since no other distribution is packaged upstream); I'll be
glad for any testing or feedback.
[1]: http://www.pbspro.org/
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
> So to sum up: it is misleading that the output from the ansible task suggests
> you to rerun with the --limit option, as it can not be used with
> elasticluster?
Yes, exactly. (But there is no way to prevent Ansible from printing
that message, AFAIK.)
> If a task fails and the cluster is
Hello Maike,
> How is this exactly supposed to be applied?
It isn't. Ansible's `--limit` option will limit also the information
gathering to the listed hosts, thus breaking ElastiCluster's playbooks
which rely on a "global view" of the cluster (i.e., much of the info
is taken dynamically from
Hello Maiken,
> fatal: [frontend001]: FAILED! => {"changed": false, "failed": true, "msg":
> "Failed to validate the SSL certificate for github.com:443. Make sure your
> managed systems have a valid CA certificate installed. You can use
> validate_certs=False if you do not need to confirm the
Dear all,
it's a pleasure to announce that ElastiCluster can now officially run
on Python3!
Many thanks to Yaroslav Halchenko for reviewing the code, suggesting
and implementing improvements.
The Docker-based installation stays based on Python 2.7 until Python 3
has received significant testing
Hello all!
I'm pleased to announce that Py3-compatible code of ElastiCluster is
now available in pull request
https://github.com/gc3-uzh-ch/elasticluster/pull/623
The code passes all tests on Travis CI for Python 3.5, 3.6, and 3.7;
and I could successfully start clusters on OpenStack, GCP and
Dear Narcis,
can you please update your ElastiCluster and try with the latest
"master" version?
The package signing key for Ubuntu had indeed changed, this is corrected in the
latest commit I pushed.
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
Dear Alex,
> TASK [lmod : Install Lmod from the OS repository (Debian-compatible)]
> **
> task path:
Dear Alex,
> TASK [glusterfs-server : include_tasks]
>
> task path:
>
Hello Alex,
what version of ElastiCluster are you running? Installed from the
Python sources or using the elasticluster.sh script / Docker
container?
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and
Dear all,
Thanks to @verdurin and his work on providing COPR repositories with
SLURM packages, SLURM 18.08 is now the default version of SLURM
installed by ElastiCluster on RHEL/CentOS 7.
Older versions are selectable with the `slurm_version` global variable.
Clusters based on Debian or Ubuntu
Hello Maiken, all,
> ERROR! 'become_user' is not a valid attribute for a TaskInclude
>
> The error appears to be in
> '/home/centos/elasticluster_20190520/src/elasticluster/share/playbooks/roles/ceph/tasks/mgr.yml':
> line 31, column 3, but may
> be elsewhere in the file depending on the exact
Hello Alex,
> How can I specify in the config file, that Elasticluster should use a
> specific availability zone for the rollout of the cluster?
Which cloud provider/backend?
Kind regards,
Riccardo
--
You received this message because you are subscribed to the Google Groups
"elasticluster"
Hi Alex,
ElastiCluster currently has no code to place nodes in a specific
availability zone on OpenStack clouds.
It would not be difficult to add it, if there is request.
I'm curious, however, as to what the use case for availability zones
would be, when using ElastiCluster. In my
Hello Alex,
many thanks for the detailed explanation! Can you please submit a
bug/feature request at
https://github.com/gc3-uzh-ch/elasticluster/issues/new ?
FWIW, we use dedicated HPC flavors (as opposed to AZs) to accomplish
the same goals on our internal OpenStack cloud.
Ciao,
R
--
You
Hello Maiken,
sorry for the late reply - I am on vacation until end of June. Can
you please open a bug at
https://github.com/gc3-uzh-ch/elasticluster/issues -- I will take a
look at it when I'm back.
Meanwhile, the workaround is to add:
safe_but_slower=yes
to your cluster's
Ah perfect I see that Manuele has solved the issue in another email thread :-)
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
Hello Gabe,
looks like a bug -- can you please open an issue report at
https://github.com/gc3-uzh-ch/elasticluster/issues ?
I am on vacation until end of June but I can take a look at it when I'm back.
Anyway: please make sure that the user or IAM profile that you are
using with ElastiCluster
Hello Maiken, all,
creation of CentOS 7.x and 6.x clusters with the default settings
(`slow_but_safer=no`, i.e., use "eatmydata" to speed up operations)
should now be fixed in the current "master" branch -- can you please
give it a try?
Ciao,
R
--
You received this message because you are
Dear all,
Nearing release 1.3.0, I would like to start discuss what directions
ElastiCluster should take in the future, what to prioritize, and in
general what the community thinks. A survey on current usage seems to
me like a good starting point to have some data to ground discussions
on.
Hello Maiken,
> I need to recreate my cluster using ipv6 only for the public adress.
Not possible at the moment -- currently, ElastiCluster playbooks
assume there is always an IPv4 address available.
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
Hello Maiken,
looks like a syntax error in the playbooks, but I won't be able to
check until end of the week.
Can you please try with Ansible 2.7 and see if the error goes away?
Ciao,
R
--
You received this message because you are subscribed to the Google Groups
"elasticluster" group.
To
Hello Maiken,
> I hacked my way through it changing the
> .elasticluster/storage/$clustername.yml file with the correct image_id for
> the nodes, but I am sure there is a better way!
Unfortunately no :-(
The way to do it without hacking the `.elasticluster/storage/*.yml`
files would be to:
Hello,
I am co-authoring a paper on different models for running "massively
parallel" experiments on the cloud; we would like to provide
additional references and examples of the use of ElastiCluster in
actual research usage (in academia or outside).
So I'm asking for help from the ElastiCluster
Hello Maiken,
this error seems to come from deep down the Paramiko dependency stack
-- can you please try with the latest ElastiCluster Docker image? It
was built yesterday night so it should have all the newest packages.
Instructions here:
Hello Maiken,
> 1.3.dev13 does not work with paramiko 2.1.1 (or 2.1.2)
To me, the error seems to come from PyOpenSSL rather than Paramiko
itself? Paramiko ends up using PyOpenSSL through the `cryptography`
Python package, so probably the version of all three packages should
be compared.
I have
Hello Maiken,
> This is because I am getting the paramiko problems with the newer version of
> elasticluster. But I should give it another try :)
I doubt that error is due to the newer ElastiCluster; sounds like a
library conflict due to (possibly) old system libraries that cannot be
updated.
Hello Maiken,
> However, I have updated my cluster to slurm 18.08.
> Can I copy over the changes wrt slurm from 1.3.dev13 to my 1.3.dev0 version?
In principle, you would only need to apply the patch in this commit:
https://github.com/gc3-uzh-ch/elasticluster/commit/20eca47
However, it's
1 - 100 of 386 matches
Mail list logo