es version : 5.2.2
- create index : index-name and index some data
ordinary index
GET index-name/_stats/?level=shards
"commit": {
"id": "jnPwdaydlNfju6jJXCUhxA==",
"generation": 8,
"user_data": {
"translog_uuid": "WJH9_RACSDK6b-Rwd-jLDA",
"sync_id": "AV91knYndKR8jYwDuHVh",
"translog_generation": "1"
},
"num_docs": 757005
}
- then shrink the index to generator a new index: read_index-name
POST index-name/_shrink/read_index-name
{
}
- index from shrink
GET read_index-name/_stats/?level=shards
"commit": {
"id": "qx4BsoFpMBjxPQWbqeaLqQ==",
"generation": 4,
"user_data": {
"translog_generation": "1",
"translog_uuid": "_6EUUgehRky_wGjIvCqHkw"
},
"num_docs": 4526847
}
- restart a node
- I got some log from es server
[2017-11-01T14:31:50,861][TRACE][o.e.i.r.PeerRecoverySourceService] [node-73-test] [index-name-20171101-000002][2] skipping [phase1] to {node-71-test}{VmZsAhrkTO6iL5jF748yGw}{wtAm1iM5S5yDjKup2VlDAQ}{10.40.2.71}{10.40.2.71:9300}{tag=hot} - identical sync id [AV91ko8tIFze9QM_1Gbj] found on both source and target
[2017-11-01T14:31:51,011][TRACE][o.e.i.r.PeerRecoverySourceService] [node-73-test] [read_index-name-20171101-000002][0] recovery [phase1] to {node-71-test}{VmZsAhrkTO6iL5jF748yGw}{wtAm1iM5S5yDjKup2VlDAQ}{10.40.2.71}{10.40.2.71:9300}{tag=hot}: not recovering [_1s.cfe], exists in local store and has checksum [1jimtv], size [405]
[2017-11-01T14:31:51,011][TRACE][o.e.i.r.PeerRecoverySourceService] [node-73-test] [read_index-name-20171101-000002][0] recovery [phase1] to {node-71-test}{VmZsAhrkTO6iL5jF748yGw}{wtAm1iM5S5yDjKup2VlDAQ}{10.40.2.71}{10.40.2.71:9300}{tag=hot}: not recovering [_1s.si], exists in local store and has checksum [1ha8n45], size [375]
[2017-11-01T14:31:51,011][TRACE][o.e.i.r.PeerRecoverySourceService] [node-73-test] [read_index-name-20171101-000002][0] recovery [phase1] to {node-71-test}{VmZsAhrkTO6iL5jF748yGw}{wtAm1iM5S5yDjKup2VlDAQ}{10.40.2.71}{10.40.2.71:9300}{tag=hot}: not recovering [_1s.cfs], exists in local store and has checksum [1kywiow], size [524352]
[2017-11-01T14:31:51,011][TRACE][o.e.i.r.PeerRecoverySourceService] [node-73-test] [read_index-name-20171101-000002][0] recovery [phase1] to
I tried some times. the index-name always recovery from local, and read_index-name always recovery from remote.
Can I to avoid the shrinked index to recovery from remove? Any advice will be great!
And how can I find a index that recovery from local or remote from CURL request , other than from es server TRACE log?