Fw: RMAN restore 128K write chunks(buffer size)
From: <Laimutis.Nedzinskas_at_seb.lt>
Date: Thu, 22 Aug 2013 16:02:44 +0300
Message-ID: <OFD68F78F7.BFDEE901-ONC2257BCF.00478756-C2257BCF.0047A9D8_at_seb.lt>
Date: Thu, 22 Aug 2013 16:02:44 +0300
Message-ID: <OFD68F78F7.BFDEE901-ONC2257BCF.00478756-C2257BCF.0047A9D8_at_seb.lt>
Case closed for now.
According to "RMAN Restore Performance from Tape is Very Poor (Doc ID 850988.1)" the alter system set "_BACKUP_KSFQ_BUFSZ"48576; should solve the issue.
brgds, laimis
Please consider the environment before printing this e-mail ----- Forwarded by Laimutis Nedzinskas/VBankas on 2013.08.22 16:01 -----
From: Laimutis Nedzinskas/VBankas To: oracle-l <oracle-l_at_freelists.org> Date: 2013.08.21 13:01 Subject: RMAN restore 128K write chunks(buffer size)
Hi,
Did anybody struggled with RMAN restore doing datafile writes in small 128K
chunks ?
128K is RMAN buffer size which is clearly a torment of underlying storage.
Metallink note 1072545.1. describes various parameters like
_backup_ksfq_bufsz and _db_file_direct_io_count but I haven't got a chance
to test them.
Brgds, Laimis N
-- http://www.freelists.org/webpage/oracle-lReceived on Thu Aug 22 2013 - 15:02:44 CEST