Error in Deepsea deployment stage.4

[Cary Reese]

Deploy ceph with Deepsea and salt-2016.11.4. openAttic 3.6.1.  Attached is image of results from deepsea monitor error on step 4/10. Final error message is Zypper command failure: Package 'python-prometheus-client' not found.

OpenAttic is deployed and seems to function so far with the exception of all default rgw.default pools are created except the data pool. Can create users / buckets, and can connect to the cluster via rgw and cloudberry client however cannot create bucket or write data. Assume due to the lack of the data pool.

Is the missing package deployment at step 4/10 responsible for this behavior and how can I install the missing python client, from where and on which node?

"ceph.monitoring.prometheus.exporters.ceph_rgw_exporter (/srv/salt/ceph/monitoring/prometheus/exporters/ceph_rgw_exporter):

  salt.ceph.local:

    install_package: An error was encountered while installing package(s): Zypper command failure: Package 'python-prometheus-client' not found."

Thanks for any help.

Cary

[volker...@googlemail.com]

Seems to be that a package repository is missing. Use one of theme here: https://build.opensuse.org/repositories/devel:languages:python

[Lenz Grimmer]

Hi Cary,

thank you for your message and sorry for the problem you've encountered.
Please note that this list is primarily about discussing
openATTIC-related issues. The DeepSea project actually has a dedicated
mailing list, that might be a better venue for resolving issues on that
layer.

Nevertheless, let's see if we can help you!

On 01/18/2018 07:21 AM, Cary Reese wrote:

> Deploy ceph with Deepsea and salt-2016.11.4. openAttic 3.6.1.  Attached
> is image of results from deepsea monitor error on step 4/10. Final error
> message is Zypper command failure: Package 'python-prometheus-client'
> not found.

Hmm, this is odd. Is this on SUSE Leap 42.3? I wonder if there has been
a change on the Open Buildservice that cause the package to disappear or
be renamed. As far as I know, 'python-prometheus-client' is required for
running the Prometheus exporter that collects object gateway statistics,
so it's not a hard dependency for openATTIC.

> OpenAttic is deployed and seems to function so far with the exception of
> all default rgw.default pools are created except the data pool. Can
> create users / buckets, and can connect to the cluster via rgw and
> cloudberry client however cannot create bucket or write data. Assume due
> to the lack of the data pool.

That might be worthwhile raising as a bug with the DeepSea team, I
wonder if there was a change related to the RGW deployment

> Is the missing package deployment at step 4/10 responsible for this
> behavior and how can I install the missing python client, from where and
> on which node?

I doubt it's related; the package is just required to monitor the
running RGW service. Due some unknown reason, DeepSea seems to have
failed to set it up correctly.

> "ceph.monitoring.prometheus.exporters.ceph_rgw_exporter
> (/srv/salt/ceph/monitoring/prometheus/exporters/ceph_rgw_exporter):
>   salt.ceph.local:
>     install_package: An error was encountered while installing
> package(s): Zypper command failure: Package 'python-prometheus-client'
> not found."

Lenz

--
SUSE Linux GmbH - Maxfeldstr. 5 - 90409 Nuernberg (Germany)
GF:Felix Imendörffer,Jane Smithard,Graham Norton,HRB 21284 (AG Nürnberg)

[Kai Wagner]

How many OSD's (disks) do you have in your test environment?

I had a problem today, that was related to rgw service deployment and to less disks during deployment because of the new introduced flag in the newest ceph version.

[Cary Reese]

Lenz,

Thank you.  The only openATTIC issue is that as it relates to the management of the rgw host, therefore the conclusion that rgw deployment was unsuccessful for some reason seems most likely correct. I’ll direct my inquiries to the Deepsea group. FYI this is using Leap 42.3 and deployment followed was the latest on openATTIC published Jan 10, 2018.  The stage.0 deployment failed in Deepsea until I installed this repository on each minion and not just on the master: http://download.opensuse.org/repositories/filesystems:openATTIC:3.x/openSUSE_Leap_42.3/filesystems:openATTIC:3.x.repo

Cary

--
You received this message because you are subscribed to a topic in the Google Groups "openATTIC Users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openattic-users/S_y6ael5k9g/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openattic-users+unsubscribe@googlegroups.com.
To post to this group, send email to openattic-users@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openattic-users/3ea4930f-4075-f571-4e28-40b1b75244f9%40suse.com.
For more options, visit https://groups.google.com/d/optout.

[Cary Reese]

VM environment with 2 OSD for each of 5 storage nodes. Had previously successfully installed an entire cluster with same vm's config with openATTIC using a trial copy of SELS 12 and ENT Storage 5, wanted to see if the open source available would provide same results.

--
You received this message because you are subscribed to a topic in the Google Groups "openATTIC Users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openattic-users/S_y6ael5k9g/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openattic-users+unsubscribe@googlegroups.com.
To post to this group, send email to openattic-users@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openattic-users/5bd061ea-5a07-7309-2de3-537abe52575c%40gmail.com.

[Kai Wagner]

I'm pretty sure that it's not possible to install and deploy with DeepSea on leap 42.3.

So if you want to test you have to have a SLE12_SP3 image for the ceph cluster.

You received this message because you are subscribed to the Google Groups "openATTIC Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openattic-use...@googlegroups.com.
To post to this group, send email to openatt...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openattic-users/CAPw54mQUyKPVyk8pyKaQ-rxD5F3Zq5vXrYib9Z2n-FBctkRqow%40mail.gmail.com.

[Denis Kondratenko]

it is actually was possible for sometime. Now it is broken with new prometheus.

https://github.com/openSUSE/vagrant-ceph/issues/11

I know patch for prometheus-node_exporter exist for deepsea:
https://github.com/SUSE/DeepSea/pull/916

but not yet for prometheus-alertmanager.

To unsubscribe from this group and all its topics, send an email to openattic-use...@googlegroups.com.

To post to this group, send email to openatt...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openattic-users/3ea4930f-4075-f571-4e28-40b1b75244f9%40suse.com.
For more options, visit https://groups.google.com/d/optout.

[Cary Reese]

The fix for me was to install the following repo…

 

#Package 'python-prometheus-client' not found. Add following repo on salt master.

 

https://download.opensuse.org/repositories/devel:/languages:/python/openSUSE_Leap_42.3/devel:languages:python.repo

 

Have a fully functional ceph cluster managed by openAttic with great metrics from Prometheus in a very very good grafana template.

 

Next project is to get NFS-Ganesha server up and running a NFS share.

 

Cary

.
To view this discussion on the web visit https://groups.google.com/d/msgid/openattic-users/864b686c-bdc6-42b6-9c69-d63ed2015762%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.