Difference between revisions of "CNM Tech Board 2023-07-20"

From CNM Wiki
Jump to: navigation, search
(Agenda)
(Proxmox)
Line 15: Line 15:
  
 
==Proxmox==
 
==Proxmox==
 +
===Old===
 +
:# The boards need to determine whether to implement CEPH or ZFS as the storage solution for ProxmoxVE
 +
:# The board needs to evaluate which servers should be included in the ProxmoxVE cluster to meet infrastructure requirements and ensuring infrastructure optimization for maximum performance.
 +
:# The board needs to determine the best method for installing Jitsi on ProxmoxVe. Options include directly integrating Jitsi with ProxmoxVE or maintaining separation by installing Jitsi on dedicated servers outside ProxmoxVe.
  
===Updates===
+
===New===
 
 
===Challenges===
 
 
 
===Proxmox===
 
 
 
 
current cluster using 3 IP for their communication channel.
 
current cluster using 3 IP for their communication channel.
 
1 ipv4 public for network cluster communication.
 
1 ipv4 public for network cluster communication.

Revision as of 17:54, 20 July 2023

CNM Technology Board 2023-7-20 is the meeting of CNM Technology Board‎ that occurred on 13 July. The predecessor meeting is CNM Technology Board 2023-7-13; the successor meeting is CNM Technology Board 2023-07-27.


  • MediaWiki
  1. The board needs to determine whether CNM MediaWiki and LDAP integration is possible.
  2. The board decided to simultaneously reach out to the previous contractor who worked on CNM MediaWiki/LDAP as well aspiring contractors who showed interest in this project to first determine whether CNM MediaWIKI/LDAP integration is possible and how much would it cost.

New businesses

  1. CEPH design. I understand that you prefer ceph for clustering. You also mentioned before that CEPH can be designed in two ways -- to combine the storage space (3 nodes * 0.5Tb = 1.5Tb) or synchronize (3 nodes, 0.5Tb each = 0.5Tb still). Which way to choose?
  1. Third server. First two servers are 0.5Tb each, but the third is 2Tb, from which, initially, I proposed to dedicate 0.5Tb to the cluster and use 1.5 outside of the cluster. Is this feasible? Or should we substitute the 2Tb (third) server with another 0.5Tb one?
  2. HA for Jitsi. Why not to use ProxmoxVE and what to use instead.
  3. Implementations. Whatever we decide, what would be needed to implement it? You mentioned "to reinstall all vms"

Open mic

Proxmox

Old

  1. The boards need to determine whether to implement CEPH or ZFS as the storage solution for ProxmoxVE
  2. The board needs to evaluate which servers should be included in the ProxmoxVE cluster to meet infrastructure requirements and ensuring infrastructure optimization for maximum performance.
  3. The board needs to determine the best method for installing Jitsi on ProxmoxVe. Options include directly integrating Jitsi with ProxmoxVE or maintaining separation by installing Jitsi on dedicated servers outside ProxmoxVe.

New

current cluster using 3 IP for their communication channel. 1 ipv4 public for network cluster communication. 2. ipv6 for network cluster communication 3. ipv4 internal for storage network cluster.

meanwhile, your new server only have 1 ipv6 and ipv4 internal. that caused, we can't connect to current cluster. there are 2 options to joining new server to the cluster possible. first remake all cluster, or second, you subscribe public ipv4 for new server.

the advantage by only using ipv6 on cluster, you can saving budget. the drawback, not all users can access ipv6.