Created attachment 1719453 [details] Current Mnimal deployment message Description of problem: ----------------------------- OCP 4.6 + OCS 4.6 Minimal deployment should not be triggered unless aggregated CPU is less than 32 or aggregated memory is less than 72 GiB. Currently it's in case aggregated CPU < 48 or memory less than 102 GiB (below the recommended config), the OCS deployment automatically falls back to minimal deployment phase. Version-Release number of selected component (if applicable): ---------------------------------------- OCP 4.6 - e.g. 4.6.0-0.nightly-2020-09-28-061045 How reproducible: ------------------- Always Steps to Reproduce: 1. Select nodes for OCS install (3 nodes preferably) ; totalCPU < 42 || totalMemory < 96; i.e 14 CPU per node(for similar node cluster) 2. UI throws warning and falls back to minimal deployment, but that is not needed. 3. Additional info: --------------------- Confirmation from Orit ``` The 42 is for a cluster of 9 osds, for a 3 osds cluster (which is the first installation) 30 cpus is enough. Minimal deployment should be triggered on if total CPU < 30 || totalMemory < 72 as it is for a 3 osds only deployment We can enforce the higher limits when the users increase the cluster capacity ```
Chat thread with detailed conversation - https://chat.google.com/room/AAAA76Pbvwc/oMLZ6LMDBMU
Created attachment 1720280 [details] screencast for the trigger Verified the following as part of this BZ: a) The trigger for minimal deployment is now changed from total(42 CPU||102 GiB) to total (30 CPU||72 GB memory)) OCP = 4.7.0-0.ci-2020-10-09-055453 OCS = ocs-operator.v4.6.0-593.ci In the screenscast ------------------------ 9 nodes(master made schedulable too), 4 CPU each = (4*9) 36 CPU if all 9 nodes selected = no warning If 8 nodes selected = 8 * 4 = 32 CPU, no warning If 7 nodes selected, 7*4 = 28 CPU <required aggregate of 30 CPU - the warning message is triggered Hence, moving this BZ to verified and the extra change requests are handled in separate BZs below: >> The warning message still needs some refinement - raised Bug 1886823 >> We still need to disable deployment if total CPU<24 and total memory<66 - raised bug Bug 1886817
*** Bug 1886823 has been marked as a duplicate of this bug. ***
https://bugzilla.redhat.com/show_bug.cgi?id=1886823 Should be fixed as part of this hence closing https://bugzilla.redhat.com/show_bug.cgi?id=1886823 as duplicate and moving it to assigned. It will be easier to send a single BZ to 4.6.z stream rather than two.
Moving it ON_QA since Nightly builds are not getting created due to some issues. (See: https://coreos.slack.com/archives/CB95J6R4N/p1603931679453000) However CI builds are working and 4.7.0-0.ci-2020-10-29-201853 contains the fix(as per the changelog)
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Moderate: OpenShift Container Platform 4.7.0 security, bug fix, and enhancement update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2020:5633