-
Notifications
You must be signed in to change notification settings - Fork 16
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added specification for expansion of cluster in tendrl post detection #254
base: master
Are you sure you want to change the base?
Conversation
@r0h4n @nthomas-redhat @julienlim @mbukatov plz check and ack. |
Please separate the detection and expand spec |
@shtripat @mbukatov @r0h4n @nthomas-redhat In reviewing 50fca31, here are some things to consider: (1) what happens if tendrl-ansible fails to get the tendrl agent(s) -- i.e. node, collectd, gluster-integration -- properly installed and configured? Can we add some mention about how this is tackled, even if it's a small mention that this is all handled at the command line. [Basically a spec should account for error-handling to some degree.] (2) can one assume all this gets logged somewhere in some tendrl log file(s) -- whether successful or failed so one can troubleshoot? (3) The "Testing" section in the document I assume refers to the "Acceptance Criteria" should include some mention that the cluster's host list in the Tendrl UI as well as Grafana will reflect the additional nodes/hosts as well as related dashboard(s). (4) Can we also mention if that this can be detection for expansion of 1 or more nodes and that is not limited to only 1 at a time, as the testing (or Acceptance Criteria) will need to allow for doing n number of nodes that get added to the Gluster trusted storage pool, i.e. work for 1 as well as work for multiple nodes. (5) Acceptance Criteria should also mention if the Inventory file is the one updated or not (as I couldn't tell if we were going with the Inventory file or with something in-memory, the alternate option), but that could be because the implementation is not yet determined. There were some minor typos but that can be easily fixed when the spec gets updated. |
@r0h4n regarding
If tendrl-node-agent not found, ideally tendrl-ansible being a command line tool should report with detailed error. I would suggest a separate specification under tendrl-ansible for this. @r0h4n ??
Anything related to tendrl-ansible should ideally with taken care as part of separate specification under tendrl-anisble. The details while import cluster flow would be updated in this specification.
Ack. I would add this acceptance criteria
Ack. I will add this acceptance criteria.
Ack. At the moment we would go ahead with inventory file and manual execution of tendrl-ansible from command line. So its like admin first modifies the inventory file for tendrl-ansible with entries for additional nodes and then runs it. This installs tendrl-node-agent on the new nodes and configures them to report to the existing tendrl server. Once this step gets over, automatic detection of nodes in tendrl server and then running import cluster flow for these additional nodes would be taken care by tendrl-server's node-agent service. I will add an acceptance criteria around inventory file of tendrl-ansible. |
@shtripat Ack your comments. I'll assume you'll make the appropriate updates to the spec. With regards to the following statement: I know if user created his/her own inventory file, probably user has to update it. However, if user is using the default /etc/ansible/hosts, should we consider adding the new entries directly there and save user the effort. This is probably a nice-to-have since we don't know yet how often users are creating his/her own inventory file or using the default Ansible hosts file. |
@julienlim agree. At this moment we expect user to manually change the inventory file and run tendrl-ansible. |
@julienlim @r0h4n added another issue #257 for specifying the detection part of new nodes in gluster cluster. @r0h4n as discussed please take care of this part of specification under a separate spec. |
50fca31
to
9e707ef
Compare
Updated the PR with more expected behavior details. |
9e707ef
to
c7468c3
Compare
tendrl-bug-id: Tendrl#253 Signed-off-by: Shubhendu <shtripat@redhat.com>
tendrl-bug-id: Tendrl#253 Signed-off-by: Shubhendu <shtripat@redhat.com>
c7468c3
to
50c7496
Compare
As per suggestions from @shtripat @r0h4n, the expansion of cluster will take place when cluster is in managed state and can be indicated as follows:
@julienlim @mcarrano @nthomas-redhat , what are you thoughts? Please update the design for the same. |
@a2batic @shtripat @nthomas-redhat @r0h4n @Tendrl/qe
Thoughts? @a2batic - We'll look into making design updates. |
@julienlim @a2batic @gnehapk During expansion for the cluster, we should show the message as Also the kebab menu options like In case there is a failure in expansion, there would be event logged from backend. Also the Thoughts?? |
Also while expansion of cluster going on the clutsrer-id should not be clickable and preferably a different icon for cluster status required. @julienlim ? |
If I read the specification correctly tendrl notice a new node automatically and via some notification requires to run tendrl-ansible, right? |
Lubos, so this spec talks about only expansion scenario and shrink part is not in the scope of this. |
@a2batic @julienlim @gnehapk @shtripat We've added a new wireframe to reflect what this should look like in the UI while the expansion is taking place. This approach can apply to expansion or any similar operation. See https://redhat.invisionapp.com/share/8QCOEVEY9#/282443505_Clusters-Update_In_Progress In summary:
Let me know if you have any questions. |
|
||
== References: | ||
|
||
* <TBD> link to tendrl-ansible issue to track the changes for this support |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm going to add a short description of this use case into README file during work on Tendrl/tendrl-ansible#82
@mbukatov @julienlim Is the view detail action(click on cluster name to view cluster detail) also allowed for expanding cluster? As per the mockup, the cluster name is hyperlinked. |
Expand cluster section added based on suggestion from fbalak. The procedure described in this commit is based on ongoing specification for cluster expansion: Tendrl/specifications#254 That said, the expand procedure is not fully finished at the time of writing this commit, see also: Tendrl/commons#849
@gnehapk @a2batic @shtripat @nthomas-redhat @r0h4n @Tendrl/qe @mcarrano Design updates have been posted at https://redhat.invisionapp.com/share/8QCOEVEY9. See Tendrl/commons#849 (comment) for further details. @gnehapk During expansion, details are still viewable. For views where there may be out-of-synch data, we show an inline notification -- see https://redhat.invisionapp.com/share/8QCOEVEY9#/screens/283595310 example. |
Signed-off-by: Shubhendu shtripat@redhat.com
tendrl-bug-id: #253