Re: Postgres 9.1 - getting a continous archiving database to accept connections taking too long

[James Sebastian <james.sebastian@xxxxxxxxx>]

Sleep interval:       60 seconds
Max wait interval:    0 forever

The whole process of looking for two history files 00000002.history and 00000001.history, 4 times and finally start the database took approximately 12 minutes.

1. How I can make this as fast to look for history files and if not present proceed quickly to start database in read/write mode ?  or not to look for history files at all.

I could get an answer in another forum explaining how sleep time of 60 seconds, ensures almost 6 minutes for waiting for WAL files and history files.

3. Each iteration, I did to see the behaviour by stopping the current postgres server, deleting the previous base backup and restoring afresh added search for one more  non existing history file like 00000004.history and 00000005.history, further slowing the process.

As I restore from the base backup each time, then push the trigger file after applying all WAL archives after the base backup, I would like to either ignore looking for history files altogether or control how many history files are tried.  A

I am still trying to figure this out on how to tell the restore process to look for only 1 or max 2 history files.

Thanks

James