利用Docker配置influxdb集群

–by leiyong 2022-4-27

docker run -d --name influxdb-server1 
-e INFLUXDB_ADMIN_USER_PASSWORD=password123 
-e INFLUXDB_USER=my_user 
-e INFLUXDB_USER_PASSWORD=my_password 
-e INFLUXDB_DB=my_database 
-e INFLUXDB_ADMIN_USER=my_admin 
-e INFLUXDB_ADMIN_USER_PASSWORD=my_password 
-e INFLUXDB_ADMIN_USER_TOKEN=token123 
--env INFLUXDB_HTTP_AUTH_ENABLED=true 
--net="bridge" 
-p 8086:8086/tcp 
-p 8088:8088/tcp 
bitnami/influxdb:1.8.5-debian-10-r257

docker run -d --name influxdb-server2 
-e INFLUXDB_ADMIN_USER_PASSWORD=password123 
-e INFLUXDB_USER=my_user 
-e INFLUXDB_USER_PASSWORD=my_password 
-e INFLUXDB_DB=my_database 
-e INFLUXDB_ADMIN_USER=my_admin 
-e INFLUXDB_ADMIN_USER_PASSWORD=my_password 
-e INFLUXDB_ADMIN_USER_TOKEN=token123 
--env INFLUXDB_HTTP_AUTH_ENABLED=true 
--net="bridge" 
-p 8090:8086/tcp 
-p 8092:8088/tcp 
bitnami/influxdb:1.8.5-debian-10-r257

syncflux用于

SyncFlux 是一个带有 HTTP API 接口的开源 InfluxDB 数据同步和复制工具，主要目标是手工从HA influxDB 1.X 集群中恢复丢失的数据

nano /mylocal/conf/syncflux.toml

# -*- toml -*-

# -------GENERAL SECTION ---------
# syncflux could work in several ways, 
# not all General config parameters works on all modes.
#  modes
#  "hamonitor" => enables syncflux as a daemon to sync 
#                2 Influx 1.X OSS db and sync data between them
#                when needed (does active monitoring )
#  "copy" => executes syncflux as a new process to copy data 
#            between master and slave databases
#  "replicashema" => executes syncflux as a new process to create 
#             the database/s and all its related retention policies 
#  "fullcopy" => does database/rp replication and after does a data copy

[General]
# ------------------------
# logdir ( only valid on hamonitor action) 
#  the directory where to place logs 
#  will place the main log "
#  

 logdir = "./log"

# ------------------------
# loglevel ( valid only for hamonitor actions ) 
#  set the log level , valid values are:
#  fatal,error,warn,info,debug,trace
# on copy/fullcopy actions  loglevel is mapped with 
#  (nothing) = Warning
#  -v = Info
#  -vv =  debug
#  -vvv = trace

 loglevel = "debug"

# -----------------------------
# sync-mode (only valid on hamonitor action)
#  NOTE: rigth now only  "onlyslave" (one way sync ) is valied
#  (planned sync in two ways in the future)

 sync-mode = "onlyslave"

# ---------------------------
# master-db choose one of the configured InfluxDB as a SlaveDB
# this parameter will be override by the command line -master parameter
 
 master-db = "influxdb01"

# ---------------------------
# slave-db choose one of the configured InfluxDB as a SlaveDB
# this parameter will be override by the command line -slave parameter
 
 slave-db = "influxdb02"

# ------------------------------
# check-interval
# the inteval for health cheking for both master and slave databases
 
 check-interval = "10s"

# ------------------------------
# min-sync-interval
# the inteval in which HA monitor will check both are ok and change
# the state of the cluster if not, making all needed recovery actions

 min-sync-interval = "20s"
 
# ---------------------------------------------
# initial-replication
# tells syncflux if needed some type of replication 
# on slave database from master database on initialize 
# (only valid on hamonitor action)
#
# none:  no replication
# schema: database and retention policies will be recreated on the slave database
# data: data for all retention policies will be replicated 
#      be carefull: this full data copy could take hours,days.
# both:  will replicate first the schema and them the full data 

 initial-replication = "none"

# 
# monitor-retry-durtion 
#
# syncflux only can begin work when master and slave database are both up, 
# if some of them is down synflux will retry infinitely each monitor-retry-duration to work.

 monitor-retry-interval = "1m"

# 
# data-chuck-duration
#
# duration for each small, read  from master -> write to slave, chuck of data
# smaller chunks of data will use less memory on the syncflux process
# and also less resources on both master and slave databases
# greater chunks of data will improve sync speed 

 data-chuck-duration = "5m"

# 
#  max-retention-interval
#
# for infinite ( or bigger ) retention policies full replication should begin somewhere in the time
# this parameter set the max retention.
 
 max-retention-interval = "8760h" # 1 year

#
#  rw-max-retries
#  
#  If any of the read ( from master) or write ( to slave ) querys fails , 
#  the query will be repeated at leas rw-max-retries
 
 rw-max-retries = 5

#  If any of the read ( from master) or write ( to slave ) querys fails , 
#  the query will be repeated at leas rw-max-retries and we can force a pause from at least rw-retry-delay
 
 rw-retry-delay = "10s"

# Num paralel  workers querying and writting at time on both databases (master & slave)
# 

 num-workers = 4

# syncflux splits  all chunk data  to write into multiple writes of max-points-on-single-write 
# enables limitation on HTTP BODY REQUEST, avoiding errors like "Request Entity Too Large"

 max-points-on-single-write = 20000

 # ---- HTTP API SECTION (Only valid on hamonitor action)
# Enables an HTTP API endpoint to check the cluster health

[http]
 name = "example-http-influxdb"
 bind-addr = "127.0.0.1:4090"
 admin-user = "admin"
 admin-passwd = "admin"
 cookie-id = "mysupercokie"

# ---- INFLUXDB  SECTION
# Sets a list of available DB's that can be used 
# as master or slaves db's on any of the posible actions

[[influxdb]]
 release = "1x"
 name = "influxdb01"
#配置集群地址、管理员名称和账号
 location = "http://192.168.83.139:8086/"
 admin-user = "my_admin"
 admin-passwd = "my_password"
 timeout = "10s"
 tls-insecure-skip-verify = true
 tls-ca = ""
 tls-cert = ""
 tls-key = ""
 
[[influxdb]]
 release = "1x"
 name = "influxdb02"
#配置集群地址、管理员名称和账号
 location = "http://192.168.83.139:8090/"
 admin-user = "my_admin"
 admin-passwd = "my_password"
 timeout = "10s"
 tls-insecure-skip-verify = true
 tls-ca = ""
 tls-cert = ""
 tls-key = ""
 

docker run -d --restart always --name=syncflux_instance00 -p 4090:4090 -v /mylocal/conf:/opt/syncflux/conf -v /mylocal/log:/opt/syncflux/log tonimoreno/syncflux

https://github.com/toni-moreno/syncflux

nano /mylocal/influxdb-srelay.conf

内容如下

###############################
##
## InfluxDB Single instances Config
##
###############################

# InfluxDB Backend InfluxDB01
[[influxdb]]
  name = "myinfluxdb01"
  location = "http://192.168.83.139:8086/"
  timeout = "10s"

# InfluxDB Backend InfluxDB02
[[influxdb]]
  name = "myinfluxdb02"
  location = "http://192.168.83.139:8090/"
  timeout = "10s"


#################################
##
## InfluxDB Cluster Configs as a set 
## of influxdb Single Instances
##
################################# 


# Cluster for linux Metrics
[[influxcluster]]
  # name = cluster id for route configs and logs
  name  = "ha_cluster"
  # members = array of influxdb backends
  members = ["myinfluxdb01","myinfluxdb02"]
  # where to write logs for all  operations in the cluster
  log-file = "ha_cluster.log"
  # log level could be
  # "panic","fatal","Error","warn","info","debug"
  log-level = "info"
  # mode = of query and send data
  # * HA : 
  #       input write data will be sent to all members
  #       queries will be sent  on the active node with 
  #       data if query-router-endpoint-api config exist, else the first
  # * Single:
  #       input write data will be sent on the first member ( should be only one)
  #     query will be sent on  the only first member (should be only one)
  # * LB:  // NOT IMPLEMENTED YET //
  type = "HA"
  # query-router-endpoint-api: 
  #	List of API url which give us the name of the influxdb backend available with all available data (when recovery process)
  # 	use any available sync tool as in https://github.com/toni-moreno/syncflux if needed 
  #
  #syncflux的配置地址
  query-router-endpoint-api = ["http://localhost:4090/api/queryactive","http://localhost:4090/api/queryactive"]


# HTTP Server
[[http]]
  name = "ha-relay"
  bind-addr = "0.0.0.0:9096"
  log-file = "http_harelay_9096.log"
  log-level = "info"
  access-log = "ha_access.log"

  rate-limit = 1000000
  burst-limit = 2000000

  ## Define endpoints base config
  ## endpoints can have read and write properties

  ## Example: /query endpoint
  ## There are 2 routes that must do a read query  against a cluster
  ## It expects that each HTTP request tries each route. If it fits the filter it will be enrouted
  ## All requests that doesn't pass through  the filter tries the next route


  #
  # IQL /query Endpoint
  #
  [[http.endpoint]]
    uri=["/query"]
    # type
    #  * RD = http for query db
    #  * WR = http for send data to the db
    type="RD"
    # source_format 
    # Supported formats
    # * ILP = influxdb-line-protocol
    # * prom-write = prometheus write format
    # * IQL = Influx Query Language
    source_format="IQL"

    ## READ request - linux_METRICS
    [[http.endpoint.route]]
      name="any_read"
      # level:
      #   * http => all following rules will work only with http params
      #   * data => any of the following rules will need data inspection
      level="http" # http or data
      # true => will use the endpoint log as this route log
      # false => will use its own log file ,  if not set the name /http_route_.log
      log-inherit = false
      #log-file = "query_route_linux_metrics.log"
      log-level = "info"

      ## Filter only the request with db param = linux_metrics
      [[http.endpoint.route.filter]]
        name="pass_all"
        #------------------------------------------------------------------------------------
        # Key for filter usage could be only at http level parameters (header/query)
        #  Header based
        #  -------------
        #    * authorization: Authorization Header 
        #    * remote-address: Remote Address Header
        #    * referer: Referer Header
        #    * user-agent: User-Agent Header
        #  Query Based 
        #  -------------
        #   (https://docs.influxdata.com/influxdb/v1.7/guides/querying_data/)
        #   (https://docs.influxdata.com/influxdb/v1.7/tools/api/#write-http-endpoint)
        #    * db [r/w]: InfluxDB to read / Write
        #    * q [r]: InfluxQL query
        #    * epoch [r]: precision on read queries
        #    * precision [w] : precision on write queries
        #    * chunked [r]: (see referenced doc)
        #    * chunksize[r]: (see referenced doc)
        #    * pretty[r]:(see referenced doc)
        #    * u [r/w]: read/write user
        #    * p [r/w]: read/write password
        #    * rp[w]: retention policy
        #    * consistency[w]: (see referenced doc)
        #  Computed
        #    * username: computed from authorization header or u parameters
        # Key for Rule Usage (not this level) could be also data level parameters
        #   * measurement: match the measurement name
        #   * tag: match the tag value with tag key in key_aux
        #   * field: match the field value with field key in key_aux (always as string!!!! at this level)
        # ----------------------------------------------------------------------------------------------
        key="db" #availabe http params
        match=".*"

      ## Rule to route to cluster_linux
      [[http.endpoint.route.rule]]
        name="route_all"
        # Action Route
        # 	* route:
        #       If key value (usually http level key) match with match parameter, The complete 
        #       request will be sent to the cluster configured in to_cluster param
        #       Next rule step will have untouched data available for any other process
	      #   * route_db_from_data (enable multitenancy)
        #       Will rename de db parameter depending for each point in data depending on
        #       the matching with one point parameter , by example one tag, enable write data
        #       to several databases (split data) from the same source.
        #       with this rule 1 HTTP request will become N HTTP request to our backends
        #       HTTP response will be logged without feedback with the original request
        #       Next rule step will have untouched data available for any other process
	      #   * rename_http
	      #   * rename_data
	      #   * drop_data
        #   * break"
        action="route"
        # See 
        key="db"
        match=".*"
        to_cluster="ha_cluster"

  [[http.endpoint]]
    uri=["/write"]
    source_format="ILP"
    type = "WR"
    
    ## WRITE request - windows
    [[http.endpoint.route]]
      name="any_write"
      level="http"

      [[http.endpoint.route.filter]]
        name="pass_all"
        key="db"
        match=".*"

      ## Send to PMEREASP15 cluster
      [[http.endpoint.route.rule]]
        name="route_all"
        action="route"
        key="db"
        match=".*"
        to_cluster="ha_cluster"

docker run -d --restart always -p 9096:9096 --volume /mylocal/influxdb-srelay.conf:/etc/influxdb-srelay/influxdb-srelay.conf     -v /mylocal/conf:/opt/influxdb-srelay/conf -v /mylocal/log:/opt/influxdb-srelay/log         tonimoreno/influxdb-srelay:latest

https://github.com/toni-moreno/influxdb-srelay

curl -l http://localhost:9096/status/ha_cluster

写入数据

curl -i -XPOST 'http://localhost:9096/write?u=my_admin&p=my_password&db=my_database' --data-binary 'cpu_load_short,host=server02 value=0.67
cpu_load_short,host=server02,region=us-west value=0.55 1422568543702900257
cpu_load_short,direction=in,host=server01,region=us-west value=2.0 1422568543702900257'

curl -G "http://localhost:9096/query?u=my_admin&p=my_password&db=my_database" --data-urlencode "q=select * from cpu_load_short"

curl https://s3.amazonaws.com/noaa.water-database/NOAA_data.txt -o NOAA_data.txt

docker cp NOAA_data.txt ead4c271ebd5:/

docker exec -it ead4c271ebd5 /bin/bash

influx -host ‘localhost’ -port ‘8086’ -username ‘my_admin’ -password ‘my_password’ -import -path=NOAA_data.txt -precision=ns