[MCOL-4762] dbroot is unassigned even in working cluster Created: 2021-06-16  Updated: 2022-10-03  Resolved: 2022-07-05

Status: Closed
Project: MariaDB ColumnStore
Component/s: cmapi
Affects Version/s: 5.5.2
Fix Version/s: cmapi-6.4.1

Type: Bug Priority: Minor
Reporter: Allen Lee (Inactive) Assignee: Roman
Resolution: Fixed Votes: 2
Labels: None

Issue Links:
PartOf
includes MCOL-4865 verification of the hostname fails , ... Closed
Sprint: 2021-9, 2021-10, 2021-11, 2021-12, 2021-13, 2021-14, 2021-15, 2021-16, 2021-17

 Description   

CMAPI status shows dbroots as empty from all nodes as below even if the cluster is just working fine.

{
"timestamp": "2021-06-15 15:52:19.092595",
"xxx.xxx.xxx.xxx": {
"timestamp": "2021-06-15 15:52:19.105190",
"uptime": 23956,
"dbrm_mode": "master",
"cluster_mode": "readwrite",
"dbroots": [],
"module_id": 1,
"services": [
{
"name": "workernode",
"pid": 27105
},
{
"name": "controllernode",
"pid": 27115
},
{
"name": "PrimProc",
"pid": 27132
},
{
"name": "ExeMgr",
"pid": 27165
},
{
"name": "WriteEngine",
"pid": 27188
},
{
"name": "DMLProc",
"pid": 27198
},
{
"name": "DDLProc",
"pid": 27229
}
]
},
.......

Is this CMAPI bug or can it be safely ignored?



 Comments   
Comment by Assen Totin (Inactive) [ 2021-06-16 ]

Constantly seeing this on a partner's CS 3-node cluster, both CMAPI 1.4 and 1.5 (the latter with a CS beta) - in fact, I have never seen CMAPI report the DB roots properly on their cluster; always just an empty list. In ColumnStore.xml everything is properly written. (I believe Allen and Roman have seen this on the cluster in question.)

The very same CS & CMAPI versions, same type of setup (3-node, GlusterFS) on my test envs always work properly. Must be something elusive, but it does happen and may cause quite a bit of confusion.

Comment by Daniel Lee (Inactive) [ 2022-06-28 ]

Build tested: 6.4.1-1 (engine #4742, cmapi #668 )

Tested both on VMs and Docker containers.

for Docker contains, mcsStatus now returns the dbroot assignment, but in format is not formatted correctly, having extra return characters.

For VM installation, dbroot assignment is still empty.

Docker installation

{
  "timestamp": "2022-06-28 02:53:30.887150",
  "mcs1": {
    "timestamp": "2022-06-28 02:53:30.892239",
    "uptime": 48056,
    "dbrm_mode": "master",
    "cluster_mode": "readwrite",
    "dbroots": [
      "1"
    ],
    "module_id": 1,
    "services": [
      {
        "name": "workernode",
        "pid": 1509
      },
      {
        "name": "controllernode",
        "pid": 1525
      },
      {
        "name": "PrimProc",
        "pid": 1536
      },
      {
        "name": "ExeMgr",
        "pid": 1573
      },
      {
        "name": "WriteEngine",
        "pid": 1586
      },
      {
        "name": "DMLProc",
        "pid": 1602
      },
      {
        "name": "DDLProc",
        "pid": 1607
      }
    ]
  },
  "mcs2": {
    "timestamp": "2022-06-28 02:53:35.915130",
    "uptime": 48061,
    "dbrm_mode": "slave",
    "cluster_mode": "readonly",
    "dbroots": [
      "2"
    ],
    "module_id": 2,
    "services": [
      {
        "name": "workernode",
        "pid": 1044
      },
      {
        "name": "PrimProc",
        "pid": 1059
      },
      {
        "name": "ExeMgr",
        "pid": 1092
      },
      {
        "name": "WriteEngine",
        "pid": 1104
      }
    ]
  },
  "mcs3": {
    "timestamp": "2022-06-28 02:53:40.935154",
    "uptime": 48066,
    "dbrm_mode": "slave",
    "cluster_mode": "readonly",
    "dbroots": [
      "3"
    ],
    "module_id": 3,
    "services": [
      {
        "name": "workernode",
        "pid": 909
      },
      {
        "name": "PrimProc",
        "pid": 923
      },
      {
        "name": "ExeMgr",
        "pid": 956
      },
      {
        "name": "WriteEngine",
        "pid": 960
      }
    ]
  },
  "num_nodes": 3
}

VM installation

[rocky8:root~]# mcsStatus
{
  "timestamp": "2022-06-28 03:16:32.435446",
  "s1pm1": {
    "timestamp": "2022-06-28 03:16:32.439030",
    "uptime": 367,
    "dbrm_mode": "master",
    "cluster_mode": "readwrite",
    "dbroots": [],
    "module_id": 1,
    "services": [
      {
        "name": "workernode",
        "pid": 7475
      },
      {
        "name": "controllernode",
        "pid": 7487
      },
      {
        "name": "PrimProc",
        "pid": 7503
      },
      {
        "name": "ExeMgr",
        "pid": 7543
      },
      {
        "name": "WriteEngine",
        "pid": 7564
      },
      {
        "name": "DMLProc",
        "pid": 7573
      },
      {
        "name": "DDLProc",
        "pid": 7602
      }
    ]
  },
  "s1pm2": {
    "timestamp": "2022-06-28 03:16:32.474683",
    "uptime": 366,
    "dbrm_mode": "slave",
    "cluster_mode": "readonly",
    "dbroots": [],
    "module_id": 2,
    "services": [
      {
        "name": "workernode",
        "pid": 6871
      },
      {
        "name": "PrimProc",
        "pid": 6892
      },
      {
        "name": "ExeMgr",
        "pid": 6932
      },
      {
        "name": "WriteEngine",
        "pid": 6950
      }
    ]
  },
  "s1pm3": {
    "timestamp": "2022-06-28 03:16:32.492050",
    "uptime": 368,
    "dbrm_mode": "slave",
    "cluster_mode": "readonly",
    "dbroots": [],
    "module_id": 3,
    "services": [
      {
        "name": "workernode",
        "pid": 6611
      },
      {
        "name": "PrimProc",
        "pid": 6632
      },
      {
        "name": "ExeMgr",
        "pid": 6672
      },
      {
        "name": "WriteEngine",
        "pid": 6684
      }
    ]
  },
  "num_nodes": 3
}

Comment by Daniel Lee (Inactive) [ 2022-07-05 ]

Build verified: cmapi 6.4.1 (b681)

dbroot number is showing in both VM and Docker installations now.

Generated at Thu Feb 08 02:52:47 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.