Regarding storage capacity when enabling the cluster via PAM

[MARUBUN SUPPORT]

Hi Team,

When enabling the cluster, the replication status did not become normal (yellow triangle).

Question 1
Is storage capacity related to the cause of the replication status not becoming normal when enabling the cluster?
From external checks, storage usage was approximately 99%.

Question 2
If there is a required storage size on the PAM server when enabling the cluster, please advise.

Thanks,

[Ralf Prigl]

Hello, Can you clarify what you mean with "storage capacity"? Is that referring to the disk usage of the individual PAM servers in the cluster, or to the capacity of the share configured for session recording? The latter should not matter for cluster synchronization, but if a PAM server runs out of disk space, it cannot download the database from the cluster master during cluster startup and will go into a bad state.

[Joseph Fry]

1: Low diskspace can absolutely prevent a node from replicating.  On cluster start, the primary site backs up its database and sends it to the secondary site node to preload the database.  If you don't have enough space on the primary node to backup the database, or enough space on the secondary site node to download the file and load it into mysql, initial replication will fail.

2.  There is no defined storage size required, as it would depend on the volume of data in the database.  Our recommended minimum for a PAM virtual appliance is 80GB.  However there are circumstances where you may need/want more.  If you suspect that something is wrong with your system that is causing excessive disk usage, please open a support case.

[MARUBUN SUPPORT]

This shows the disk usage for each individual server in the cluster.

Just to confirm, it's understood that the storage size for each individual server in the cluster doesn't need to be the same, correct?
However, once synchronization completes, the usage will become the same.

Also, have you heard of any instances where AWS disk sizes were expanded?

[Ralf Prigl]

Hello, the disk size doesn't need to be the same for nodes in a cluster, but obviously it makes sense to use a common size that is appropriate for your deployment. The documented minimum disk/storage size is 80GB, and that is adequate for the vast majority of PAM implementations. As Joe Fry mentioned in his update, if you find that you have high disk usage with an 80GB disk, you should engage PAM Support to investigate what is using up the disk space rather than look into increasing disk space right away.

The disk usage will not be the same on all cluster nodes. Each nodes writes its own log files. The replication leader tends to write more logs than other nodes and thus have somewhat larger disk usage, but the difference would be expected to be just a few percent. User and API activity also will impact local log sizes. How well user and API traffic is distributed across the cluster depends on your specific implementation, including your implementation of load balancing.  Problems with writing database backups, session recordings, or archiving of credential manager metric and auditlog data can cause increased disk usage. These all relate to local configurations that can be different on each node.

[MARUBUN SUPPORT]

When expanding storage with an AWS PAM instance, is it correct to assume that no processing needs to be done on the PAM side?
I tried an experiment and it took a while to expand, but after the expansion was complete I ran PAM and checked the dashboard, etc., and no particular errors were found.

[Ralf Prigl]

Hello, PAM encrypts its disk so that the data is protected against retrieval by someone who manages to mount the disk on another server. When you expand the disk, PAM needs to format/encrypt the added space, which takes some time. If your question is about whether you need to take any action from the PAM UI afterwards, the answer is No.