Murmur 1.2.19 unable to recover from timeout on DB connection with DB on HA solution #1632

New issue

Closed

opened 2026-02-20 21:08:45 -05:00 by deekerman · 3 comments

deekerman commented

2026-02-20 21:08:45 -05:00

Owner

Originally created by @Glowsome on GitHub (Sep 15, 2019).

Situation :

Murmur 1.2.19 running on server 1
MySQL backend running on Server 2 , situated on HA cluster VM (Proxmox V6.x) with shared storage.
Database correctly configured for murmur, and functional.

Behaviour :

Migrate MySQL backend server to another node (usually takes like a few seconds)
Murmur daemon receives a timeout on the database connection while migration is in progress
murmur daemon never recovers from the connection loss, daemon keeps running but throws error 'unable to connect on database' and will not allow clients to connect to server untill server daemon is manually reloaded/restarted.

Suggestions to solve :

Implement an option for 'cluster-awareness' regarding database, increasing timeout option for connection to database.
increase time before throwing a fatal to reconnect database ( in conjunction to cluster-awareness)
buffer database info in murmur to not loose functionality while a/the database is migrating in the background (or possibly in a non-clustered env a/the database-server is restarting)

Remark :
As this is experienced on version 1.2.9 and we are now at new release 1.3.0 but have not migrated to it due to another bug, so i have yet to investigate the behavior on 1.3.0

Originally created by @Glowsome on GitHub (Sep 15, 2019). Situation : - Murmur 1.2.19 running on server 1 - MySQL backend running on Server 2 , situated on HA cluster VM (Proxmox V6.x) with shared storage. - Database correctly configured for murmur, and functional. Behaviour : - Migrate MySQL backend server to another node (usually takes like a few seconds) - Murmur daemon receives a timeout on the database connection while migration is in progress - murmur daemon never recovers from the connection loss, daemon keeps running but throws error 'unable to connect on database' and will not allow clients to connect to server untill server daemon is manually reloaded/restarted. Suggestions to solve : - Implement an option for 'cluster-awareness' regarding database, increasing timeout option for connection to database. - increase time before throwing a fatal to reconnect database ( in conjunction to cluster-awareness) - buffer database info in murmur to not loose functionality while a/the database is migrating in the background (or possibly in a non-clustered env a/the database-server is restarting) Remark : As this is experienced on version 1.2.9 and we are now at new release 1.3.0 but have not migrated to it due to another bug, so i have yet to investigate the behavior on 1.3.0

deekerman

2026-02-20 21:08:45 -05:00

closed this issue
added the
bug

server
labels

deekerman commented

2026-02-20 21:08:48 -05:00

Author

Owner

@Krzmbrzl commented on GitHub (Jun 6, 2020):

As this is experienced on version 1.2.9 and we are now at new release 1.3.0 but have not migrated to it due to another bug, so i have yet to investigate the behavior on 1.3.0

Did you migrate yet? If so: does the problem persist?

@Krzmbrzl commented on GitHub (Jun 6, 2020): > As this is experienced on version 1.2.9 and we are now at new release 1.3.0 but have not migrated to it due to another bug, so i have yet to investigate the behavior on 1.3.0 Did you migrate yet? If so: does the problem persist?

deekerman commented

2026-02-20 21:08:48 -05:00

Author

Owner

@Glowsome commented on GitHub (Jun 6, 2020):

@Krzmbrzl , the same issue exists/is seen on in 1.3.0

@Glowsome commented on GitHub (Jun 6, 2020): @Krzmbrzl , the same issue exists/is seen on in 1.3.0

deekerman commented

2026-02-20 21:08:48 -05:00

Author

Owner

@Krzmbrzl commented on GitHub (Jun 6, 2020):