Test Run - Insufficient Memory?

[mpgxvii]

Hi @K1Hyve @afolarin @blootsvoets
I am finally able to run most components by installing them by batches. Previously, I was getting timeout errors from different components (sometimes upload connector, sometimes others). To solve that, I've been setting install: false to some components first then setting them to true later, then running helmfile sync again.

NAME                                             READY   STATUS             RESTARTS   AGE
catalog-server-7d759fddd7-dszhf                  0/1     Init:0/1           1          62m
cp-kafka-0                                       2/2     Running            0          41m
cp-kafka-1                                       2/2     Running            0          40m
cp-kafka-2                                       2/2     Running            0          39m
cp-kafka-rest-6bcfbcbf99-km2pw                   0/2     Pending            0          32m
cp-schema-registry-dc77c5975-pxqzv               2/2     Running            0          34m
cp-zookeeper-0                                   2/2     Running            0          107m
cp-zookeeper-1                                   2/2     Running            0          105m
cp-zookeeper-2                                   2/2     Running            0          104m
management-portal-68b5b5cc95-smzm2               1/1     Running            5          55m
minio-0                                          0/1     Pending            0          21m
minio-1                                          0/1     Pending            0          21m
minio-2                                          0/1     Pending            0          21m
minio-3                                          0/1     Pending            0          21m
minio-update-prometheus-secret-tmtkg             0/1     Pending            0          21m
postgresql-postgresql-0                          3/3     Running            0          55m
radar-fitbit-connector-56664f869c-67hkl          1/1     Running            9          54m
radar-grafana-79d5455f97-wxvdn                   1/1     Running            0          50m
radar-jdbc-connector-5d5dbb6d4f-hl4bd            1/1     Running            8          49m
radar-output-5b7f79d656-j5lf7                    0/1     CrashLoopBackOff   8          21m
radar-rest-sources-authorizer-678b7f7b87-hgbpk   1/1     Running            0          54m
radar-rest-sources-backend-8d585f56-bkln8        1/1     Running            0          50m
radar-s3-connector-8d6649b59-vrr6r               0/1     Pending            0          4m7s
radar-s3-connector-d5dd686cf-p9dpz               0/1     Pending            0          21m
radar-upload-connect-backend-7b588cb7cc-8btqd    1/1     Running            0          44m
radar-upload-connect-frontend-66cc6cb95c-tpxfz   1/1     Running            0          43m
radar-upload-postgresql-postgresql-0             3/3     Running            0          49m
redis-master-0                                   2/2     Running            0          21m
s3-proxy-58f4d88cd6-jbh5m                        1/1     Running            0          21m
timescaledb-postgresql-0                         3/3     Running            0          50m

Now some components just stay in Pending state and I'm getting when I describe the minio pod:

  Type     Reason            Age                    From               Message
  ----     ------            ----                   ----               -------
  Warning  FailedScheduling  7m38s (x2 over 7m38s)  default-scheduler  0/1 nodes are available: 1 pod has unbound immediate PersistentVolumeClaims.
  Warning  FailedScheduling  8s (x7 over 7m35s)     default-scheduler  0/1 nodes are available: 1 Insufficient memory.

What's the minimum amount of memory needed? Can I make changes to the values so I can run everything on my local minikube instance (on Macbook Pro)?

[keyvaann]

If you have installed monitoring and logging systems you need around 32gb of memory. You can disable those components to reduce the memory requirement.
You can also reduce the value of request.memory of the charts to allow installation.

[afolarin]

A list of known minimum resources for e.g. core, monitoring etc... would be useful for the README. It's the kind of stuff people won't know a priori.

[blootsvoets]

Made a PR to include minimum specs #163.

[keyvaann]

PR is merged now.

[mpgxvii]

Thanks! I disabled the monitoring components but still getting similar error. But changing the memory requests for some components seems to make it work.

[keyvaann]

You're welcome, there's is also logging systems that you can disable to reduce memory usage.