Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[0.11.0-rc1] Cluster with -join need ALL node to restart #6027

Closed
PierreF opened this issue Mar 16, 2016 · 2 comments
Closed

[0.11.0-rc1] Cluster with -join need ALL node to restart #6027

PierreF opened this issue Mar 16, 2016 · 2 comments

Comments

@PierreF
Copy link
Contributor

PierreF commented Mar 16, 2016

I've seen #5602, it's great! With the addition of #5706 it now possible to start cluster on every node with just:

$ influxd -join node1:8091,node2:8091,node3:8091 -hostname $HOSTNAME

As I would expect (and from my understanding of PR #5602), to restart node I could still keep the SAME command line; e.g. :

$ influxd -join node1:8091,node2:8091,node3:8091 -hostname $HOSTNAME

This works well only if at least one node was still alive. If all node are down (let's say a power outage), if only two node are back online (let's say of node3 had an hardware fault following the power outage), I got the following:

2016/03/16 14:26:02 InfluxDB starting, version 0.11.0rc1, branch 0.11, commit 441772e87782c27a679043071f7181f7928bfbb2
2016/03/16 14:26:02 Go version go1.4.3, GOMAXPROCS set to 4
2016/03/16 14:26:02 no configuration provided, using default settings
[meta] 2016/03/16 14:26:02 Starting meta service
[meta] 2016/03/16 14:26:02 Listening on HTTP: [::]:8091
[metastore] 2016/03/16 14:26:02 Using data dir: /root/.influxdb/meta
[metastore] 2016/03/16 14:26:02 Waiting for 2 join peers.  Have [node1:8088]. Asking nodes: [node1:8091 node2:8091 node3:8091]
[metastore] 2016/03/16 14:26:03 Waiting for 1 join peers.  Have [node1:8088 node2:8088]. Asking nodes: [node1:8091 node2:8091 node3:8091]
[...]
[metastore] 2016/03/16 14:27:08 Waiting for 1 join peers.  Have [node1:8088 node2:8088]. Asking nodes: [node1:8091 node2:8091 node3:8091]

It hang forever :( I would expect cluster to restart as soon as quorum is reached. I need to have ALL three node online to restart the cluster.

It seem to be close to #5494, shouldn't we ignore -join argument if node is already a member of a cluster ?

@PierreF
Copy link
Contributor Author

PierreF commented Mar 22, 2016

Any updates ? It would be very useful for us to have this fixed before 0.11 release.

@pauldix
Copy link
Member

pauldix commented Mar 24, 2016

0.11 was released today. Future versions of the open source InfluxDB will be standalone servers. See https://influxdata.com/blog/update-on-influxdb-clustering-high-availability-and-monetization/

@pauldix pauldix closed this as completed Mar 24, 2016
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants