Creating custom retention policy on 0.9

[prashleo]

I am working on proposing the right configuration of influxdb 0.9 to go with our product and the performance statistics of using influxDB 0.9

I was able to successfully setup the 3 node cluster and they are synched(creating a database on one node is reflected on all the nodes)

But when I insert measurements into the database on one node, it is not replicated on the other two This is because it is using the default retention policy ( which has the replication set to 1)

I am trying to create a custom retention policy like this and I get this error.

It is a 3 node cluster and I am using a replication of 3: appreciate your kind help

curl -G 'http://10.91.39.35:8086/query'; --data-urlencode "q=CREATE retention policy test on testdb duration 1h replication 3"

{"results":[{"error":"exec failed: apply: replication factor must match cluster size; this limitation will be lifted in v0.9.1"}]}

Steps used to create the 3 node cluster: 1-> download and install the latest 0.9 influxdb on all the 3 nodes 2-> generate the influxdb config file and update the paths (from /root/.influxdb/... to var/opt/influxdb/...) due to permission issue

3. change the hostname to actual ip address and add the peers[ip1:8088, ip2:8088, ip3:8088] under [meta]
4. start the influxdb server on all three nodes

[otoolep]

This error implies that your cluster is not actually coming together, since the node believes it's only 1 node. Do the logs on each node indicate that a cluster has formed?

[prashleo]

2015/06/24 19:46:46 [DEBUG] raft: Node 10.91.38.223:8088 updated peer set (2): [127.0.0.1:8088] 2015/06/24 19:46:46 [DEBUG] raft: Node 10.91.38.223:8088 updated peer set (2): [127.0.0.1:8088] 2015/06/24 19:46:46 [DEBUG] raft: Node 10.91.38.223:8088 updated peer set (2): [10.91.38.210:8088 10.91.39.35:8088 10.91.38.223:8088]

similar message is there on all the 3 nodes, and also when I create a Database one one node, it shows up on all other nodes as wel

[prashleo]

also there is no "cluster" tab on the web UI as it used to be in 0.88??

[prashleo]

Carefully redoing the 3 node setup solved this for me, again having the cluster tab in the GUI was so helpful, atleast to verify that all the nodes are coming together. thanks otoolep

[otoolep]

OK, glad you sorted it out.

[prashleo]

Thanks for your help otoolep.

I have an issue inserting multiple measurements, its failing with "bad time stamp" but the same time stamp works perfect when inserting just one measurement/points

is it the write syntax( I got this syntax from Influxdb document)

lab@influxdb0:~$ curl -i -XPOST ' http://10.91.38.210:8086/write?db=testdb&rp=test&precision=s' -d 'cpu_load_short,host=server01,region=us-west value=0.64 cpu_load_short,host=server02,region=us-west value=0.55 1435281408 cpu_load_short1,direction=in,host=server01,region=us-west value=23422.0 1435281408'

HTTP/1.1 400 Bad Request

Request-Id: 6d76f70b-1ba1-11e5-8013-000000000000

X-Influxdb-Version: FIXME

Date: Fri, 26 Jun 2015 01:19:44 GMT

Content-Length: 13

Content-Type: text/plain; charset=utf-8

bad timestamplab@influxdb0:~$ -Prashanth On Thu, Jun 25, 2015 at 5:05 PM, otoolep notifications@github.com wrote:

OK, glad you sorted it out.

— Reply to this email directly or view it on GitHub https://github.com/influxdb/influxdb/issues/3133#issuecomment-115434484.

[otoolep]

We can take a look, but can you open a new issue first?

[prashleo]

done https://github.com/influxdb/influxdb/issues/3147

On Thu, Jun 25, 2015 at 6:41 PM, otoolep notifications@github.com wrote:

We can take a look, but can you open a new issue first?

— Reply to this email directly or view it on GitHub https://github.com/influxdb/influxdb/issues/3133#issuecomment-115454106.