What folders to use with Folder2Ram within a Cluster + Ceph environment to minimize disk wear out

[logui]

I have a Proxmox cluster with 3 nodes + Ceph enabled, no HA. I am trying to optimize the writing of logs to disk (SSD), to minimize its degradation over time due to excessive log writing to the SSD. I have implemented Folder2Ram initially with the following folders :

• /var/log
• /var/lib/pve-cluster
• /var/lib/pve-manager
• /var/lib/rrdcached

I think with these folders I am addressing most of the PVE Cluster logging into RAM, but I might be missing some of the Ceph logging folders, should I add something else? Thanks

 

[LnxBil]

Just buy proper small enterprise SSDs for a few bucks and don't concern yourself with a crash recovery if your have a cluster cold start and /var/lib/pve-cluster is not available or outdated.

 

[fiona]

Hi,

I have a Proxmox cluster with 3 nodes + Ceph enabled, no HA. I am trying to optimize the writing of logs to disk (SSD), to minimize its degradation over time due to excessive log writing to the SSD. I have implemented Folder2Ram initially with the following folders :

• /var/log
• /var/lib/pve-cluster
• /var/lib/pve-manager
• /var/lib/rrdcached

I think with these folders I am addressing most of the PVE Cluster logging into RAM, but I might be missing some of the Ceph logging folders, should I add something else? Thanks

you'll likely get issues down the line: https://forum.proxmox.com/threads/p...s-persis-after-host-reboot.135121/post-597558

 

[logui]

Thank you, I found that most of the writing to the disk is coming from the Ceph Monitors (ceph-mon) vs journald, now, I am trying to find a way to send them to memory or disable them:

• ceph-mon -f --cluster ceph --id N3 --setuser ceph --setgroup ceph [rocksdb:low]
• ceph-mon -f --cluster ceph --id N3 --setuser ceph --setgroup ceph [ms_dispatch]

 

[TheRealMaN_]

Thank you, I found that most of the writing to the disk is coming from the Ceph Monitors (ceph-mon) vs journald, now, I am trying to find a way to send them to memory or disable them:

• ceph-mon -f --cluster ceph --id N3 --setuser ceph --setgroup ceph [rocksdb:low]
• ceph-mon -f --cluster ceph --id N3 --setuser ceph --setgroup ceph [ms_dispatch]

Did you find a solution?