Hi ,

I got this error while doing my online backups on the new postgres
8.0.2. I follow the same procedure as mentioned in the docs and the way
i do in brief is

1) Put the database in backup mode with pg_start_backup
2) copy data cluster, I use tablespaces and my tables and indexes are in
different partition so i copy them also here using tar and gzip
3) And put the database back in normal mode with pg_stop_backup
4) copy archive log directory with tar and gzip.

and today I go this error , i am not sure if this is normal or is
this something to worry about ? does anybody experienced anything like
this before ?

ERROR Durring Backup
--------------------------------------------------------------------------------
bash-2.05b$ ./backup.sh b
17/8E3DA9F0
tar: tabledata/165899758/165958430: file changed as we read it
tar: Error exit delayed from previous errors
17/8E8CB0A8
bash-2.05b$
----------------------------------------------------------------------------------

Thanks!
Pallav


---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend

Pallav Kalva wrote:
> Hi ,
>
> I got this error while doing my online backups on the new postgres
> 8.0.2. I follow the same procedure as mentioned in the docs and the way
> i do in brief is
>
> 1) Put the database in backup mode with pg_start_backup
> 2) copy data cluster, I use tablespaces and my tables and indexes are in
> different partition so i copy them also here using tar and gzip
> 3) And put the database back in normal mode with pg_stop_backup
> 4) copy archive log directory with tar and gzip.
>
> and today I go this error , i am not sure if this is normal or is
> this something to worry about ? does anybody experienced anything like
> this before ?
>
> ERROR Durring Backup
> --------------------------------------------------------------------------------
> bash-2.05b$ ./backup.sh b
> 17/8E3DA9F0
> tar: tabledata/165899758/165958430: file changed as we read it
> tar: Error exit delayed from previous errors
> 17/8E8CB0A8
> bash-2.05b$
> ----------------------------------------------------------------------------------

This is expected. Tar is backing up a changing file system but the
WAL logging will bring the system back to a consistent state. However,
read the current version of the docs online because you have to be sure
you have the WAL files that were active during the backup to use the tar
file, and those files might still be sitting in pg_xlog and not archived
at the time the tar finishes.

--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

Pallav Kalva <pkalva@deg.cc> writes:
> bash-2.05b$ ./backup.sh b
> 17/8E3DA9F0
> tar: tabledata/165899758/165958430: file changed as we read it
> tar: Error exit delayed from previous errors
> 17/8E8CB0A8
> bash-2.05b$

You may need to use a stupider version of tar that doesn't consider that
an error condition. (Offhand I see nothing in the gnu tar man page
about suppressing this error check :-() Perhaps cpio would do instead?

The problem is not that you have an unusable backup, it is that if your
script ignores error exits from tar, you won't be able to detect real
problems.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend