add gathering scenario sql

oceanbase · Jul 16, 2024 · 57a6dd3 · 57a6dd3
1 parent faac571
commit 57a6dd3
Show file tree

Hide file tree

Showing 11 changed files with 56 additions and 2 deletions.
diff --git a/handler/gather/tasks/observer/clog_disk_full.yaml b/handler/gather/tasks/observer/clog_disk_full.yaml
@@ -105,6 +105,15 @@ task:
       - type: sql
         sql: "select * from oceanbase.DBA_OB_SERVER_EVENT_HISTORY where event like '%migrat%' and name6 like '%fail%' and value6=1;"
         global: true
+      - type: sql
+        sql: "select b.zone, a.svr_ip,a.svr_port,b.with_rootserver, round(a.total_size/1024/1024/1024,0) as total_GB,round(a.free_size/1024/1024/1024,0) as free_GB, round((a.total_size-a.free_size)/1024/1024/1024,0) as used_GB,(a.total_size-a.free_size)/a.total_size as used_percent from oceanbase.__all_virtual_disk_stat a, oceanbase.DBA_OB_SERVERS b where a.svr_ip=b.svr_ip and a.svr_port=b.svr_port order by b.zone,used_percent desc;"
+        global: true
+      - type: sql
+        sql: "select tenant_name, svr_ip,  memstore_limit /(1024 * 1024 * 1024) as memstore_limit_GB,  freeze_trigger /(1024 * 1024 * 1024) as freeze_trigger_GB,  memstore_used /(1024 * 1024 * 1024) as memstore_used_GB,  concat((memstore_used * 100 / memstore_limit), '%') as memstore_used_percent,  active_span /(1024 * 1024 * 1024) as active_span_GB,  freeze_cnt   from oceanbase.GV$OB_MEMSTORE memstore_info  inner join oceanbase.DBA_OB_TENANTS tenant on  memstore_info.tenant_id = tenant.tenant_id  ORDER BY tenant.tenant_name,svr_ip;"
+        global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
+        global: true
       - type: ssh
         ssh: "df -h"
         global: false

diff --git a/handler/gather/tasks/observer/compaction.yaml b/handler/gather/tasks/observer/compaction.yaml
@@ -150,6 +150,15 @@ task:
       - type: sql
         sql: "select t.tenant_name, t1.database_name, round(sum(t2.data_size)/1024/1024/1024,2) as data_size_gb, round(sum(t2.required_size)/1024/1024/1024,2) as required_size_gb from oceanbase.dba_ob_tenants t, oceanbase.cdb_ob_table_locations t1, oceanbase.cdb_ob_tablet_replicas t2 where t.tenant_id=t1.tenant_id and t1.svr_ip=t2.svr_ip and t1.tenant_id=t2.tenant_id and t1.ls_id=t2.ls_id and t1.tablet_id=t2.tablet_id and t1.role='leader' group by t.tenant_name, t1.database_name order by data_size_gb desc;"
         global: true
+      - type: sql
+        sql: "select b.zone, a.svr_ip,a.svr_port,b.with_rootserver, round(a.total_size/1024/1024/1024,0) as total_GB,round(a.free_size/1024/1024/1024,0) as free_GB, round((a.total_size-a.free_size)/1024/1024/1024,0) as used_GB,(a.total_size-a.free_size)/a.total_size as used_percent from oceanbase.__all_virtual_disk_stat a, oceanbase.DBA_OB_SERVERS b where a.svr_ip=b.svr_ip and a.svr_port=b.svr_port order by b.zone,used_percent desc;"
+        global: true
+      - type: sql
+        sql: "select tenant_name, svr_ip,  memstore_limit /(1024 * 1024 * 1024) as memstore_limit_GB,  freeze_trigger /(1024 * 1024 * 1024) as freeze_trigger_GB,  memstore_used /(1024 * 1024 * 1024) as memstore_used_GB,  concat((memstore_used * 100 / memstore_limit), '%') as memstore_used_percent,  active_span /(1024 * 1024 * 1024) as active_span_GB,  freeze_cnt   from oceanbase.GV$OB_MEMSTORE memstore_info  inner join oceanbase.DBA_OB_TENANTS tenant on  memstore_info.tenant_id = tenant.tenant_id  ORDER BY tenant.tenant_name,svr_ip;"
+        global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
+        global: true
       - type: log
         global: false
         grep: ""
diff --git a/handler/gather/tasks/observer/delay_of_primary_and_backup.yaml b/handler/gather/tasks/observer/delay_of_primary_and_backup.yaml
@@ -156,6 +156,12 @@ task:
       - type: sql
         sql: "SELECT LS_ID, SCN_TO_TIMESTAMP(END_SCN) FROM oceanbase.GV$OB_LOG_STAT WHERE ROLE = 'LEADER';"
         global: true
+      - type: sql
+        sql: "select b.zone, a.svr_ip,a.svr_port,b.with_rootserver, round(a.total_size/1024/1024/1024,0) as total_GB,round(a.free_size/1024/1024/1024,0) as free_GB, round((a.total_size-a.free_size)/1024/1024/1024,0) as used_GB,(a.total_size-a.free_size)/a.total_size as used_percent from oceanbase.__all_virtual_disk_stat a, oceanbase.DBA_OB_SERVERS b where a.svr_ip=b.svr_ip and a.svr_port=b.svr_port order by b.zone,used_percent desc;"
+        global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
+        global: true
       - type: log
         global: false
         grep: ""
diff --git a/handler/gather/tasks/observer/log_archive.yaml b/handler/gather/tasks/observer/log_archive.yaml
@@ -120,6 +120,9 @@ task:
       - type: sql
         sql: "SELECT * FROM oceanbase.CDB_OB_ARCHIVELOG_PIECE_FILES limit 20"
         global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
+        global: true
       - type: log
         global: false
         grep: ""

diff --git a/handler/gather/tasks/observer/long_transaction.yaml b/handler/gather/tasks/observer/long_transaction.yaml
@@ -84,6 +84,9 @@ task:
       - type: sql
         sql: "SELECT count(1) FROM oceanbase.GV$OB_TRANSACTION_PARTICIPANTS WHERE CTX_CREATE_TIME < date_sub(now(), INTERVAL 600 SECOND) AND STATE = 'INIT';"
         global: true
+      - type: sql
+        sql: "select tenant_name, svr_ip,  memstore_limit /(1024 * 1024 * 1024) as memstore_limit_GB,  freeze_trigger /(1024 * 1024 * 1024) as freeze_trigger_GB,  memstore_used /(1024 * 1024 * 1024) as memstore_used_GB,  concat((memstore_used * 100 / memstore_limit), '%') as memstore_used_percent,  active_span /(1024 * 1024 * 1024) as active_span_GB,  freeze_cnt   from oceanbase.GV$OB_MEMSTORE memstore_info  inner join oceanbase.DBA_OB_TENANTS tenant on  memstore_info.tenant_id = tenant.tenant_id  ORDER BY tenant.tenant_name,svr_ip;"
+        global: true
       - type: log
         grep: ""
         global: false
diff --git a/handler/gather/tasks/observer/memory.yaml b/handler/gather/tasks/observer/memory.yaml
@@ -94,7 +94,10 @@ task:
         sql: "show parameters like '%syslog_io_bandwidth_limit%';"
         global: true
       - type: sql
-        sql: "select * from oceanbase.GV$OB_MEMSTORE limit 20"
+        sql: "select tenant_name, svr_ip,  memstore_limit /(1024 * 1024 * 1024) as memstore_limit_GB,  freeze_trigger /(1024 * 1024 * 1024) as freeze_trigger_GB,  memstore_used /(1024 * 1024 * 1024) as memstore_used_GB,  concat((memstore_used * 100 / memstore_limit), '%') as memstore_used_percent,  active_span /(1024 * 1024 * 1024) as active_span_GB,  freeze_cnt   from oceanbase.GV$OB_MEMSTORE memstore_info  inner join oceanbase.DBA_OB_TENANTS tenant on  memstore_info.tenant_id = tenant.tenant_id  ORDER BY tenant.tenant_name,svr_ip;"
+        global: true
+      - type: sql # 大于1g的内存模块
+        sql: "SELECT CASE WHEN tenant_name IS NULL THEN TENANT_ID ELSE tenant_name END tenant_name, host,ctx_name, mod_name, hold, used, count FROM ( SELECT tenant_name,tenant_id,HOST,ctx_name,mod_name,hold,used,COUNT, ROW_NUMBER () OVER ( PARTITION BY tenant_name, HOST ORDER BY hold desc) rnum FROM   (SELECT  b.tenant_name, a.tenant_id, concat(a.svr_ip, ':', a.svr_port) HOST,  a.ctx_name, a.mod_name,  round(a.hold / 1024 / 1024 / 1024) hold, round(a.used / 1024 / 1024 / 1024) used, a.COUNT FROM  oceanbase.__all_virtual_memory_info a  LEFT JOIN oceanbase.__all_tenant b ON a.TENANT_ID = b.TENANT_ID  WHERE a.hold > 1024 * 1024 * 1024 ));"
         global: true
       - type: ssh # 可看到租户的规格、线程、队列及请求统计等信息，且这条日志每个租户每 30s 打印一次
         ssh: "grep 'dump tenant info.tenant=' ${observer_data_dir}/log/observer.log | sed 's/,/,/g'"

diff --git a/handler/gather/tasks/observer/recovery.yaml b/handler/gather/tasks/observer/recovery.yaml
@@ -89,6 +89,9 @@ task:
       - type: sql
         sql: "SELECT * FROM oceanbase.CDB_OB_RESTORE_HISTORY limit 20;"
         global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
+        global: true
       - type: log
         grep: ""
         global: false

diff --git a/handler/gather/tasks/observer/rootservice_switch.yaml b/handler/gather/tasks/observer/rootservice_switch.yaml
@@ -132,6 +132,9 @@ task:
       - type: sql
         sql: "SELECT TENANT_NAME, TENANT_ID, TENANT_ROLE, STATUS, SWITCHOVER_STATUS FROM oceanbase.DBA_OB_TENANTS"
         global: true
+      - type: sql # 大于1g的内存模块
+        sql: "SELECT CASE WHEN tenant_name IS NULL THEN TENANT_ID ELSE tenant_name END tenant_name, host,ctx_name, mod_name, hold, used, count FROM ( SELECT tenant_name,tenant_id,HOST,ctx_name,mod_name,hold,used,COUNT, ROW_NUMBER () OVER ( PARTITION BY tenant_name, HOST ORDER BY hold desc) rnum FROM   (SELECT  b.tenant_name, a.tenant_id, concat(a.svr_ip, ':', a.svr_port) HOST,  a.ctx_name, a.mod_name,  round(a.hold / 1024 / 1024 / 1024) hold, round(a.used / 1024 / 1024 / 1024) used, a.COUNT FROM  oceanbase.__all_virtual_memory_info a  LEFT JOIN oceanbase.__all_tenant b ON a.TENANT_ID = b.TENANT_ID  WHERE a.hold > 1024 * 1024 * 1024 ));"
+        global: true
       - type: log
         grep: ""
         global: false

diff --git a/handler/gather/tasks/observer/suspend_transaction.yaml b/handler/gather/tasks/observer/suspend_transaction.yaml
@@ -81,6 +81,9 @@ task:
       - type: sql
         sql: "SELECT count(1) FROM oceanbase.GV$OB_TRANSACTION_PARTICIPANTS WHERE CTX_CREATE_TIME < date_sub(now(), INTERVAL 600 SECOND) AND (STATE = 'PREPARE' OR STATE = 'REDO COMPLETE' OR STATE ='PRECOMMIT');"
         global: true
+      - type: sql
+        sql: "select tenant_name, svr_ip,  memstore_limit /(1024 * 1024 * 1024) as memstore_limit_GB,  freeze_trigger /(1024 * 1024 * 1024) as freeze_trigger_GB,  memstore_used /(1024 * 1024 * 1024) as memstore_used_GB,  concat((memstore_used * 100 / memstore_limit), '%') as memstore_used_percent,  active_span /(1024 * 1024 * 1024) as active_span_GB,  freeze_cnt   from oceanbase.GV$OB_MEMSTORE memstore_info  inner join oceanbase.DBA_OB_TENANTS tenant on  memstore_info.tenant_id = tenant.tenant_id  ORDER BY tenant.tenant_name,svr_ip;"
+        global: true
       - type: log
         grep: ""
         global: false
diff --git a/handler/gather/tasks/observer/unit_data_imbalance.yaml b/handler/gather/tasks/observer/unit_data_imbalance.yaml
@@ -154,7 +154,10 @@ task:
         sql: "select t.tenant_name, t1.database_name, round(sum(t2.data_size)/1024/1024/1024,2) as data_size_gb, round(sum(t2.required_size)/1024/1024/1024,2) as required_size_gb from oceanbase.dba_ob_tenants t, oceanbase.cdb_ob_table_locations t1, oceanbase.cdb_ob_tablet_replicas t2 where t.tenant_id=t1.tenant_id and t1.svr_ip=t2.svr_ip and t1.tenant_id=t2.tenant_id and t1.ls_id=t2.ls_id and t1.tablet_id=t2.tablet_id and t1.role='leader' group by t.tenant_name, t1.database_name order by data_size_gb desc;"
         global: true
       - type: sql
-        sql: "select svr_ip,total_size / 1024 / 1024 / 1024 total_G,free_size / 1024 / 1024 / 1024 free_G,(total_size - free_size) / 1024 / 1024 / 1024 used_G,(total_size - free_size) / total_size used_percentage FROM oceanbase.__all_virtual_disk_stat;  "
+        sql: "select b.zone, a.svr_ip,a.svr_port,b.with_rootserver, round(a.total_size/1024/1024/1024,0) as total_GB,round(a.free_size/1024/1024/1024,0) as free_GB, round((a.total_size-a.free_size)/1024/1024/1024,0) as used_GB,(a.total_size-a.free_size)/a.total_size as used_percent from oceanbase.__all_virtual_disk_stat a, oceanbase.DBA_OB_SERVERS b where a.svr_ip=b.svr_ip and a.svr_port=b.svr_port order by b.zone,used_percent desc;"
+        global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
         global: true
       - type: log
         grep: ""

diff --git a/handler/gather/tasks/observer/unknown.yaml b/handler/gather/tasks/observer/unknown.yaml
@@ -69,6 +69,15 @@ task:
       - type: sql
         sql: "SELECT a.TENANT_NAME,a.TENANT_ID,b.SVR_IP FROM oceanbase.DBA_OB_TENANTS a, oceanbase.GV$OB_UNITS b WHERE a.TENANT_ID=b.TENANT_ID;"
         global: true
+      - type: sql
+        sql: "select b.zone, a.svr_ip,a.svr_port,b.with_rootserver, round(a.total_size/1024/1024/1024,0) as total_GB,round(a.free_size/1024/1024/1024,0) as free_GB, round((a.total_size-a.free_size)/1024/1024/1024,0) as used_GB,(a.total_size-a.free_size)/a.total_size as used_percent from oceanbase.__all_virtual_disk_stat a, oceanbase.DBA_OB_SERVERS b where a.svr_ip=b.svr_ip and a.svr_port=b.svr_port order by b.zone,used_percent desc;"
+        global: true
+      - type: sql # 大于1g的内存模块
+        sql: "SELECT CASE WHEN tenant_name IS NULL THEN TENANT_ID ELSE tenant_name END tenant_name, host,ctx_name, mod_name, hold, used, count FROM ( SELECT tenant_name,tenant_id,HOST,ctx_name,mod_name,hold,used,COUNT, ROW_NUMBER () OVER ( PARTITION BY tenant_name, HOST ORDER BY hold desc) rnum FROM   (SELECT  b.tenant_name, a.tenant_id, concat(a.svr_ip, ':', a.svr_port) HOST,  a.ctx_name, a.mod_name,  round(a.hold / 1024 / 1024 / 1024) hold, round(a.used / 1024 / 1024 / 1024) used, a.COUNT FROM  oceanbase.__all_virtual_memory_info a  LEFT JOIN oceanbase.__all_tenant b ON a.TENANT_ID = b.TENANT_ID  WHERE a.hold > 1024 * 1024 * 1024 ));"
+        global: true
+      - type: sql # 查看每台机器上的tablet分布
+        sql: "select svr_ip, svr_port, ls_id, count(*) from oceanbase.CDB_OB_TABLE_LOCATIONS group by svr_ip, svr_port, ls_id;"
+        global: true
       - type: sql
         sql: "show parameters like '%syslog_level%';"
         global: true