> On Feb 11, 2015, at 4:12 AM, Daniel <[email protected]> wrote:
> 
> 
> 
> - create a recovery.conf file as per the wal-e info page.
>   restore_command = 'envdir /etc/wal-e.d/env wal-e wal-fetch "%f" "%p"'
>   
>   I've tried with and without a recovery_target_time equal to the time of the 
> backup and it doesn't change things.
>   
> - start postgres
> 
> Postgres won't start, here are the logs:
> 
> Feb 11 10:53:10 vagrant postgres[26779]: [2-1] 2015-02-11 10:53:10 GMT   LOG: 
>  database system was interrupted; last known up at 2015-02-11 04:45:01 GMT
> Feb 11 10:53:10 vagrant postgres[26779]: [3-1] 2015-02-11 10:53:10 GMT   LOG: 
>  starting point-in-time recovery to 2015-02-11 04:45:01+00
> Feb 11 10:53:10 vagrant postgres[26779]: [4-1] 2015-02-11 10:53:10 GMT   LOG: 
>  invalid checkpoint record
> Feb 11 10:53:10 vagrant postgres[26779]: [5-1] 2015-02-11 10:53:10 GMT   
> FATAL:  could not locate required checkpoint record
> Feb 11 10:53:10 vagrant postgres[26779]: [5-2] 2015-02-11 10:53:10 GMT   
> HINT:  If you are not restoring from a backup, try removing the file 
> "/var/lib/postgresql/9.3/main/backup_label".
> Feb 11 10:53:10 vagrant postgres[26778]: [2-1] 2015-02-11 10:53:10 GMT   LOG: 
>  startup process (PID 26779) exited with exit code 1
> Feb 11 10:53:10 vagrant postgres[26778]: [3-1] 2015-02-11 10:53:10 GMT   LOG: 
>  aborting startup due to startup process failure
> 
> If I do exactly the same procedure with two vagrant VMs (ie backup one to s3 
> and restore the other) everything goes swimmingly well, I'm using the same S3 
> keys, buckets, policies etc.
> 
> Looking around, most comments on "invalid checkpoint record" seem to be to do 
> with WAL files not getting archived during the backup, however as far as I 
> can see that's ok.  I assume that if postgres called the recovery_command to 
> fetch a wal backup and that failed, I'd see the call to wal-e in the logs.
> 
> I'd be grateful if anyone could help, or even just give me a more detailed 
> description of what 'invalid checkpoint record' means.

Based on the fact that your log indicates no WAL files were restored, I’d guess 
your wal-fetch is failing for some reason.

Usually wal-e will log some stuff to syslog.  Can you look in /var/log/syslog 
and see if there are any failure reports and hopefully a helpful error?


-- 
You received this message because you are subscribed to the Google Groups 
"wal-e" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to