View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000770 | Pgpool-II | Bug | public | 2022-10-01 18:11 | 2023-01-19 22:50 |
| Reporter | rblst | Assigned To | |||
| Priority | high | Severity | minor | Reproducibility | sometimes |
| Status | new | Resolution | open | ||
| Platform | Linux | OS | Red Hat Enterprise Linux | OS Version | 8.3 |
| Product Version | 4.0.19 | ||||
| Summary | 0000770: kind mismatch among backends, kind details are: 0[2] 1[Z] 2[Z] messages for very simple SELECTs | ||||
| Description | The following type of messages appear in pgpool logs for simple SELECTs. Sep 13 13:12:06 pgpool-db01tk pgpool[769072]: 2022-09-30 13:12:06: pid 2009452: WARNING: packet kind of backend 0 ['2'] does not match with master/majority nodes packet kind ['Z'] Sep 13 13:12:06 pgpool-db01tk pgpool[769072]: 2022-09-30 13:12:06: pid 2009452: FATAL: failed to read kind from backend Sep 13 13:12:06 pgpool-db01tk pgpool[769072]: 2022-09-30 13:12:06: pid 2009452: DETAIL: kind mismatch among backends. Possible last query was: "select excel_0.status as col_0 from gadget.excel excel_0" kind details are: 0[2] 1[Z] 2[Z] Sep 13 13:12:06 pgpool-db01tk pgpool[769072]: 2022-09-30 13:12:06: pid 2009452: HINT: check data consistency among db nodes | ||||
| Steps To Reproduce | Streaming replication (master/slave) mode is used Java PostgreSQL driver is used | ||||
| Tags | No tags attached. | ||||
|
|
I cannot reproduce the problem here. I think I need more info. Can you share following info? - pgpool.conf - pgpool log with log_client_messages enabled |
|
|
Hi, any update on this error? |
|
|
I am still waiting for the info. > I cannot reproduce the problem here. I think I need more info. Can you share following info? > - pgpool.conf > - pgpool log with log_client_messages enabled |
|
|
Hi, unfortunately, I have no direct access to the server, plus it is probably no longer in production. I will try once again to have the admins turn on log_client_messages and try to reproduce the errors for a usable log, but no guarantees, sorry. |
