we need a system to do crontab tasks that beanstalkd may works ,but this system is asked to be high available and can't loss service ability while it's turned down or just killed.
Any plan on HA cluster like replication ? Or any advice to implement HA ?
JodeZer
All 3 comments
Given the current implementation, you could always enable WALing and make sure your init process restarts beanstalkd if it crashes.
If you want restarts beyond a single machine, you could mount the same NFS on another machine and run something like consul lock beanstalkd ... on all HA VMs.
JensRantil
on 2 Mar 2017
@JensRantil Thanks a lot!
etcd or ZK ways happened to have been discussed not long ago.
JodeZer
on 3 Mar 2017
Another approach is to to 1) make sure your tasks are idempotent and 2) do double writes to two beanstalkd instances.
I'll be closing this. Let me know if you still have outstanding issues.
JensRantil
on 19 Aug 2018
Related issues
ramSeraph
路
4Comments
kr
路
4Comments
ysmolsky
路
10Comments
ysmolsky
路
7Comments
raju-divakaran
路
3Comments
Most helpful comment
Given the current implementation, you could always enable WALing and make sure your init process restarts beanstalkd if it crashes.
If you want restarts beyond a single machine, you could mount the same NFS on another machine and run something like
consul lock beanstalkd ...on all HA VMs.