[LU-17819] LMR1a: Replicate MGS Services to Multiple MDS

• Type: Technical task
• Resolution: Unresolved
• Priority: Minor
• Assignee: Lai Siyao
• Reporter: Andreas Dilger

Description

Replicate MGS running on MDT0000 to other MDTs/MDSes to improve MGS availability and resiliency in the face of server and/or storage issues. This includes:

• redundant MGS/MGT configuration logs
• redundant MGS Imperative Recovery service

Issue Links

• is cloned by: LU-18823 LMR1b: Replicate MDT0000 Services to Other MDS
• is related to: LU-19046 server mount stuck uninterruptibly in mgc_fs_setup()
• is related to: LU-19199 Ensure all fast recovery features are enabled and working
• is related to: LU-16722 MGS config log restructuring
• is related to: LU-19584 FTM: port raft to kernel
• is related to: LU-19585 FTM: Implement Raft in Lustre
• is related to: LU-19591 FTM: Sync configs to other MDTs from leader MDT with MGS
• is related to: LU-19592 FTM: set permanent parameters by Raft protocol
• is related to: LU-19593 FTM: Processing Nodemap in RAFT
• is related to: LU-19594 FTM: support IR on replicated MGS
• is related to: LU-19595 FTM: support 64bit index llog
• is related to: LU-19316 mgs_target_reg() code refactoring