Bug 2142016
| Summary: | Service is ready after upscale even when upscale is not finished | ||
|---|---|---|---|
| Product: | [Red Hat Storage] Red Hat OpenShift Data Foundation | Reporter: | Filip Balák <fbalak> |
| Component: | odf-managed-service | Assignee: | Leela Venkaiah Gangavarapu <lgangava> |
| Status: | CLOSED WONTFIX | QA Contact: | Oded <oviner> |
| Severity: | high | Docs Contact: | |
| Priority: | low | ||
| Version: | 4.10 | CC: | aeyal, dbindra, ebenahar, lgangava, nberry, ocs-bugs, odf-bz-bot, oviner |
| Target Milestone: | --- | Keywords: | TestBlocker |
| Target Release: | --- | Flags: | fbalak:
needinfo?
(nberry) fbalak: needinfo? (oviner) |
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2023-06-05 10:56:21 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
> State is ready even when upscale is in progress.
- OLM will reconcile resources in 8 minute intervals afaik and it'll take this much time for the current status to be surfaced
- However, with 8 minutes unless there are any blocking issues every component will be reconciled during the upscale and unless the OLM timing matches the upscale, user will not see the flip in status of service
- Due to above this can be a RFE if at all even possible, but will be of lower priority even then.
This means that there is no observability for the feature which should be a valid bug and not a RFE. @Neha, @Elad WDYT? Hi Leela, There is no way that the described behavior is expected. Even if the fix is to come with an RFE, this issue still needs to be considered a bug - ack, I'll take a re-look if our deployer can partially surface the status - However, OCS is thinking that when the user asked for 12TiB and cluster is running with 4TiB isn't a failure scenario and so isn't changing the status of the operand (storagecluster cr) because eventually it'll be fulfilled - @dbindra since you've a fair grasp on OCS is this something you could check? Current Scenario: 1. Cluster is running with 4TiB 2. User scaled it to 12TiB 3. There's a window b/n the size is scaled from 4TiB to 12TiB and it might so happen that due to OCM new nodes will not come online 4. During this window new OSD pods will be in pending state however OCS doesn't see that as a potential failure scenario thanks, leela. Once an addon is installed then the addon can either go into failed state or uninstalled state There is no updating/progressing state after successful installation so we cannot surface the status if the customer is upscaling the cluster I would suggest having this as an RFE and forwarding the request to OCM While upscaling does the status of storageCluster change to Progressing? > While upscaling does the status of storageCluster change to Progressing?
- yes, the storagecluster status will be in progressing status and so the deployer state
- however, within 8 min the progressing will be turned to Ready unless there's an issue and so OCM will not see the transient state
@fbalak could you pls re-check just after the upscale, what's the status of "storagecluster", "managedocs" CRs and "rosa list services"
- pls note, 4TiB to 8TiB will be very quick and so would suggest scale from 4TiB to 20TiB and this gives some time for observation of status in above resources and possibly in rosa cli as well
@oviner Would you please turn this bug into RFE and provide information requested by Leela? |
Description of problem: If user performs upscale: $ rosa edit service id=2HOOmEi3QEvWvCp5rpRev9ss3SL -size="8" I: Service "2HOOmEi3QEvWvCp5rpRev9ss3SL" is now updating. To check the status run 'rosa describe service --id 2HOOmEi3QEvWvCp5rpRev9ss3SL' then there is recommended command "rosa describe service" in output but if the command is executed then Service state is ready and there is no indication that anything is in progress even when upscale is not finished. $ rosa describe service --id 2HOOmEi3QEvWvCp5rpRev9ss3SL Id: 2HOOmEi3QEvWvCp5rpRev9ss3SL Href: /api/service_mgmt/v1/services/2HOOmEi3QEvWvCp5rpRev9ss3SL Service type: ocs-provider-dev Service State: ready Cluster Name: fbalak-pr Created At: 2022-11-11 07:46:59 +0000 UTC Updated At: 2022-11-11 11:21:10 +0000 UTC Parameters: "onboarding-validation-key" : "<key>" "notification-email-1" : "fbalak" "notification-email-2" : "odf-ms-qe" "size" : "8" Version-Release number of selected component (if applicable): ocs-osd-deployer.v2.0.8 How reproducible: 2/2 Steps to Reproduce: 1. Upscale cluster: rosa edit service --id=<service_ID> --size="<new_size in TiB>" 2. Verify that upscale finished: $ rosa describe service --id <service_ID> Actual results: State is ready even when upscale is in progress. Expected results: User is informed that upscale is in progress. Additional info: