Обсуждение: BUG #13645: pg_basebackup backup hash index & unlogged table
The following bug has been logged on the website: Bug reference: 13645 Logged by: digoal Email address: digoal@126.com PostgreSQL version: 9.4.4 Operating system: CentOS 6.x x64 Description: HI, Hash index and unlogged table don't generate XLOG record, when we use pg_basebackup's backup datafile & xlog archive recovery database, unlogged table datafile will purged, and hash index also has unconsistent data page. Why pg_basebackup not filter the hash index datafile & unlogged table datafile, these file waste net band and backup file's capcity.
On Mon, Sep 28, 2015 at 06:15:24AM +0000, digoal@126.com wrote: > The following bug has been logged on the website: > > Bug reference: 13645 > Logged by: digoal > Email address: digoal@126.com > PostgreSQL version: 9.4.4 > Operating system: CentOS 6.x x64 > Description: > > HI, > Hash index and unlogged table don't generate XLOG record, when we use > pg_basebackup's backup datafile & xlog archive recovery database, unlogged > table datafile will purged, and hash index also has unconsistent data page. > Why pg_basebackup not filter the hash index datafile & unlogged table > datafile, these file waste net band and backup file's capcity. We will issue a warning in 9.5 when hash indexes are created. I am not sure why we don't filter out hash indexed and unlogged tables. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + As you are, so once was I. As I am, so you will be. + + Roman grave inscription +
On Mon, Oct 5, 2015 at 10:37 AM, Bruce Momjian <bruce@momjian.us> wrote: > On Mon, Sep 28, 2015 at 06:15:24AM +0000, digoal@126.com wrote: >> Hash index and unlogged table don't generate XLOG record, when we use >> pg_basebackup's backup datafile & xlog archive recovery database, unlogged >> table datafile will purged, and hash index also has unconsistent data page. >> Why pg_basebackup not filter the hash index datafile & unlogged table >> datafile, these file waste net band and backup file's capcity. > > We will issue a warning in 9.5 when hash indexes are created. I am not > sure why we don't filter out hash indexed and unlogged tables. basebackup.c does not have much logic yet to see if the file it is working on is a relation file (this already exists in filemap.c@isRelDataFile in pg_rewind), and a WAL sender is not connected to a particular database, which would be needed for pg_class lookups, so it seems to me that this would be a complicated infrastructure for a niche use case. Note to mention that perhaps, in a perhaps-distant day, hash indexes will have WAL support. -- Michael