[Improvement] Default size of /dev/shm is too small for PostgreSQL

[nayutah]

Is your improvement request related to a problem? Please describe.
Default size of /dev/shm in docker & containerd is 64MB, too small for databases relying on shared memory like PostgreSQL
Also referred to bug #6013

The problem has been solved in:
https://github.com/apecloud/kubeblocks-addons/blob/main/addons/apecloud-postgresql/templates/clusterdefinition.yaml

One can test it with kubeblocks 0.7.* series:
slc@slcmac aws % kubectl exec -it poplar07-postgresql-0 -- /bin/sh
Defaulted container "postgresql" out of: postgresql, pgbouncer, metrics, kb-checkrole, config-manager, pg-init-container (init)
# df -h
Filesystem Size Used Avail Use% Mounted on
overlay 457G 128G 307G 30% /
tmpfs 64M 0 64M 0% /dev
/dev/sda2 457G 128G 307G 30% /etc/hosts
tmpfs 126G 8.0K 126G 1% /kb-podinfo
tmpfs 126G 40K 126G 1% /dev/shm
tmpfs 126G 12K 126G 1% /run/secrets/kubernetes.io/serviceaccount
tmpfs 63G 0 63G 0% /proc/acpi
tmpfs 63G 0 63G 0% /proc/scsi
tmpfs 63G 0 63G 0% /sys/firmware

root@node5:~# crictl exec -it 2b9e18c1cd576 /bin/sh
# df -h
Filesystem Size Used Avail Use% Mounted on
overlay 457G 47G 387G 11% /
tmpfs 64M 0 64M 0% /dev
/dev/sda2 457G 47G 387G 11% /etc/hosts
tmpfs 126G 8.0K 126G 1% /kb-podinfo
tmpfs 126G 40K 126G 1% /dev/shm
tmpfs 126G 12K 126G 1% /run/secrets/kubernetes.io/serviceaccount
tmpfs 63G 0 63G 0% /proc/acpi
tmpfs 63G 0 63G 0% /proc/scsi
tmpfs 63G 0 63G 0% /sys/firmware

This solution is adopted by several PostgreSQL providers, but this solution also has a safety risk:
When you dont set the 'sizeLimit', the default size of /dev/shm is set up with 'node physical memory capacity'.
It is a dangerous behavior for containers & k8s, a malicious container can congest all node memory with:
dd if=/dev/zero of=/dev/shm/output bs=4096 count=1M (or other counts)
So other containers on the same node will be affected.

So it is better to :
1, explicitly set 'sizeLimit' in the pod spec
2, if 'sizeLimit' is unset, the Operator should give it a default value with pod memory 'request' @leon-inf