Bug 2160294 - [GSS] Procedure to increase OSD pod memory limits
Summary: [GSS] Procedure to increase OSD pod memory limits
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat OpenShift Data Foundation
Classification: Red Hat Storage
Component: ocs-operator
Version: 4.8
Hardware: All
OS: All
high
high
Target Milestone: ---
: ---
Assignee: Nitin Goyal
QA Contact: Elad
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2023-01-12 00:53 UTC by kelwhite
Modified: 2023-09-12 08:02 UTC (History)
13 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2023-08-10 02:43:12 UTC
Embargoed:

Attachments (Terms of Use)

Description kelwhite 2023-01-12 00:53:27 UTC 
Description of problem (please be detailed as possible and provide log
snippests):

We need a way to increase an OSD pods memory limits. Since the osd is managed by a replica set which is managed by a deployment which is managed by the cephcluster which is managed by the storagecluster, you *should* be able to run this patch command to update the OSD pod memory limits (it works for the MDS pods!)

# oc patch StorageCluster ocs-storagecluster --type='merge' -p '{"spec":{"resources":{"osd":{"limits":{"cpu":"3","memory":"16Gi"},"requests":{"cpu":"3","memory":"16Gi"}}}}}'

Furthermore, if this is not the correct way to update the OSD pod memory limits, what is? Should this be updated via the rook-config-override CM?
Comment 3 kelwhite 2023-01-12 18:02:49 UTC 
Seems the limits are defined under storageDeviceSets and not under spec: resources:

# example:
apiVersion: ocs.openshift.io/v1    
kind: StorageCluster    
metadata:    
  namespace: openshift-storage    
  name: example-storagecluster    
spec:
  manageNodes: false    
  monPVCTemplate:    
    spec:    
      storageClassName: gp2    
      accessModes:    
      - ReadWriteOnce    
      resources:    
        requests:    
          storage: 1Gi    
  resources:      
    mon:      
        resources:
            requests:
                cpu:
                memory:
            limits:
                cpu:
                memory:  
    mds:      
        resources:
            requests:
                cpu:
                memory:
            limits:
                cpu:
                memory:     
    rgw:      
        resources:
            requests:
                cpu:
                memory:
            limits:
                cpu:
                memory:     
    mgr:      
        resources:
            requests:
                cpu:
                memory:
            limits:
                cpu:
                memory:   
    noobaa-core:      
        resources:
            requests:
                cpu:
                memory:
            limits:
                cpu:
                memory:     
    noobaa-db:        
        resources:
            requests:
                cpu:
                memory:
            limits:
                cpu:
                memory:    
  storageDeviceSets:    
  - name: example-deviceset    
    count: 3    
    resources:
      requests:
        cpu:
        memory:
      limits:
        cpu:
        memory: 
    placement: {}    
    dataPVCTemplate:    
      spec:    
        storageClassName: gp2    
        accessModes:    
        - ReadWriteOnce    
        volumeMode: Block    
        resources:    
          requests:    
            storage: 100Gi    
    portable: true

Seems we just need a patch command to update the storageDeviceSets, something like the following?:

# oc patch StorageCluster ocs-storagecluster --type='merge' -p '{"spec":{"storageDeviceSets":{"resources":{"cpu":"3","memory":"16Gi"},"requests":{"cpu":"3","memory":"16Gi"}}}}}'
Note You need to log in before you can comment on or make changes to this bug.