Description
As part of clustered monitoring I am having a look, on a 3 n1ql node cluster, at
localhost:8093/admin/clusters
worryingly, it reports
[{"name":"default","datastore":"http://127.0.0.1:8091","configstore":"http://127.0.0.1:8091","accountstore":"","version":""}]
|
even though the node name is actually 172.23.123.103
If I drill down
localhost:8093/admin/clusters/default/nodes
[{"cluster":"default","name":"172.23.123.103","queryEndpoint":"http://172.23.123.103:8093/query/service","adminEndpoint":"http://172.23.123.103:8093/admin","querySecure":"https://172.23.123.103:18093/query/service","adminSecure":"https://172.23.123.103:18093/admin","options":null}]
|
it only reports the first node (at least it's named correctly)
However - having a look at /pools/default, I do see all nodes and they have N1QL listed.
..."hostname":"172.23.123.105:8091","clusterCompatibility":327680,"version":"5.0.0-1841-enterprise","os":"x86_64-unknown-linux-gnu","ports":{"sslProxy":11214,"httpsMgmt":18091,"httpsCAPI":18092,"proxy":11211,"direct":11210},"services":["index","kv","n1ql"]}
|
seems like clustering/couchbase/clustering_cb.go:QueryNodeNames does not collect all the n1ql nodes...
Attachments
For Gerrit Dashboard: MB-22588 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
73015,3 | MB-20773 make sure that planner selects #nodes index It took me the best part of a day to debug this by instrumentation. The planner wasn't selecting the #nodes index because the system indexer would only ever return the primary index. | master | query | Status: MERGED | +2 | +1 |
73322,3 | MB-22588 Fix the error in spawning spatial updater | master | couchdb | Status: MERGED | +2 | +1 |