In the storage-computing separation deployment mode, why does one of the three nodes have no disk space?

24 views Asked by At

CnosDB version: enterprise 2.3.4. Result of sql SHOW DATANODES;.

+---------------------+---------------------------------------+-----------+-------------+-----------+----------+---------------------+
| NODE_ID             | HOST                                  | ATTRIBUTE | STATUS      | DISK_FREE | LOCATION | LAST_UPDATED_TIME   |
+---------------------+---------------------------------------+-----------+-------------+-----------+----------+---------------------+
| 1703747750800040865 | cnosdbpro-tskv-2.cnosdbpro-tskv.admin | HOT       | NODISKSPACE | 0 B       |          | 2024-01-29 06:20:09 |
| 1703747750800040866 | cnosdbpro-tskv-1.cnosdbpro-tskv.admin | HOT       | HEALTHY     | 2.02 GB   |          | 2024-01-29 06:20:06 |
| 1703747750800040867 | cnosdbpro-tskv-0.cnosdbpro-tskv.admin | HOT       | HEALTHY     | 2.02 GB   |          | 2024-01-29 06:20:07 |
+---------------------+---------------------------------------+-----------+-------------+-----------+----------+---------------------+. 

It can be seen that while both nodes still have 2GB of space, one node has no space. Why is this, shouldn't writes to storage nodes be even?

config:

host = "cnosdbpro-tskv-2.cnosdbpro-tskv.admin"
license_file = "/etc/cnosdb/license.json"
reserve_space = "0"
[cache]
[cluster]
  flight_rpc_listen_port = 8904
  grpc_listen_port = 8903
  http_listen_port = 8902
  meta_service_addr = ["cnosdbpro-meta-0.cnosdbpro-meta.admin:8901", "cnosdbpro-meta-1.cnosdbpro-meta.admin:8901", "cnosdbpro-meta-2.cnosdbpro-meta.admin:8901"]
  name = "cnosdbpro"
  tcp_listen_port = 8905
  vector_listen_port = 8906
[deployment]
[heartbeat]
  report_time_interval_secs = 30
[hinted_off]
  enable = true
  path = "/var/lib/cnosdb/hh"
[log]
  level = "info"
  path = "/var/log/cnosdb"
[node_basic]
  cold_data_server = false
  location = ""
  node_id = 1703747750800040865
  store_metrics = true
[query]
  auth_enabled = false
  max_server_connections = 10240
  query_sql_limit = 16777216
  read_timeout_ms = 3000
  stream_executor_cpu = 2
  stream_trigger_cpu = 1
  write_sql_limit = 167772160
  write_timeout_ms = 3000
[security]
[storage]
  path = "/var/lib/cnosdb/data"
  strict_write = false
[subscription]
  cache = 1028
  concurrency = 8
  timeout = 300
[wal]
  enabled = true
  path = "/var/lib/cnosdb/wal"

The above is my config file. I deployed it through the operator, so most of the parameters are default and not much modification was made.

0

There are 0 answers