Hello,
I've created synchronous replication between Primary and Secondary server and put pacemaker + PAF over. Client is doing some stress tests (switching nodes, disabling particular node, etc) and it's second time when this kind of problem occoures :
2020-12-18 14:03:46.658 CET [unknown] [28787]ERROR: requested starting point F/A2000000 on timeline 39 is not in this server's history
2020-12-18 14:03:46.658 CET [unknown] [28787]DETAIL: This server's history forked from timeline 39 at F/A1023338.
2020-12-18 14:03:46.658 CET [unknown] [28787]DETAIL: This server's history forked from timeline 39 at F/A1023338.
questions:
1. what does this mean ? How this can happen ? Does this mean that for some "point in time" both servers were primary ?
2. in xxx.history files I find the following rows:
43 F/A60000A0 no recovery target specified
44 F/A70000A0 no recovery target specified
45 F/A80000A0 no recovery target specified
again: what does this mean ?
3. general question: can anyone suggest deep explanation of timelines and history files to understand how this postgresql nature works ?
43 F/A60000A0 no recovery target specified
44 F/A70000A0 no recovery target specified
45 F/A80000A0 no recovery target specified
again: what does this mean ?
3. general question: can anyone suggest deep explanation of timelines and history files to understand how this postgresql nature works ?
Many thx
Marcin
