Database Apps Suspended After Region Transfer on Fly.io - How to Restart Without Data Loss?

16 Views Asked by At

I've recently encountered an issue with my applications hosted on Fly.io. Initially, all my applications were running smoothly in the Chennai (MAA) region. However, I received a notification from Fly.io stating that the MAA region would be deprecated and my applications would be transferred to the Mumbai (BOM) region.

On Monday, following the region transfer, my database applications were suspended unexpectedly. The logs indicate a system restart followed by a series of actions leading to the shutdown of various processes, including the database server, proxy, and supervisor processes (logs provided below for reference). Notably, timestamps in the logs are missing, displaying "Invalid date".

Invalid date    
[5313503.257433] reboot: Restarting system
Invalid date    
Umounting /dev/vdb from /data
Invalid date    
Starting clean up.
Invalid date    
keeper   | Process exited 0
Invalid date    
keeper   | server stopped
Invalid date    
keeper   |  done
Invalid date    
keeper   | waiting for server to shut down....2024-03-25 17:45:46.925 UTC [588] LOG:  database system is shut down
Invalid date    
keeper   | 2024-03-25 17:45:46.916 UTC [590] LOG:  shutting down
Invalid date    
keeper   | 2024-03-25 17:45:46.916 UTC [588] LOG:  background worker "logical replication launcher" (PID 595) exited with exit code 1
Invalid date    
keeper   | 2024-03-25 17:45:46.913 UTC [588] LOG:  aborting any active transactions
Invalid date    
proxy    | exit status 130
Invalid date    
keeper   | 2024-03-25 17:45:46.911 UTC [588] LOG:  received fast shutdown request
Invalid date    
proxy    | [WARNING] 084/174546 (538) : All workers exited. Exiting... (130)
Invalid date    
proxy    | [ALERT] 084/174546 (538) : Current worker #1 (561) exited with code 130 (Interrupt)
Invalid date    
proxy    | [NOTICE] 084/174546 (538) : path to executable is /usr/sbin/haproxy
Invalid date    
proxy    | [NOTICE] 084/174546 (538) : haproxy version is 2.2.9-2+deb11u3
Invalid date    
sentinel | Process exited 0
Invalid date    
exporter | signal: interrupt
Invalid date    
proxy    | [WARNING] 084/174546 (538) : Exiting Master process...
Invalid date    
proxy    | Stopping interrupt...
Invalid date    
sentinel | Stopping interrupt...
Invalid date    
keeper   | Stopping interrupt...
Invalid date    
exporter | Stopping interrupt...
Invalid date    
supervisor stopping
Invalid date    
Got terminated, stopping
Invalid date    
Sending signal SIGTERM to main child process w/ PID 521

Upon investigation, I noticed that while my volume appears to be intact in the MAA region, there are no machines listed for my application. This has left me unable to restart my applications and I am particularly concerned about potential data loss.

Has anyone faced a similar situation, or can anyone provide guidance on how to restart my applications in the BOM region without losing any data? Any help or pointers towards relevant documentation or steps to follow would be greatly appreciated.

0

There are 0 best solutions below