1. parameter explain BACKUP_TAPE_IO_SLAVES
If BACKUP_TAPE_IO_SLAVES=TRUE a single tape slave per channel is used.
This speeds up the backup if the bottleneck is in writing to tape;
it frees the channel process to continue processing rman input buffers
whilst the tape slave waits for IO completion.
2. run backup job when backup_tape_io_slaves set true
SQL> /
INST SID SERIAL# USERNAME PROGRAM MODULE MACHINE CLIENT_INFO SQL_ID B_INST B_SESS EVENT S_WAIT SPID
----- ----- ---------- ---------- ------------------------------ ------------------------------ --------------- -------------------- --------------- ------- ------- --------------------------------------------- ---------- ------------------------
1 8 5 SYS sqlplus@orcldb (TNS V1-V3) sqlplus@orcldb (TNS V1-V3) orcldb 6pduj1hhyxg0f SQL*Net message to client 0 20258
1 11 9 SYS i/o slave wait 0 21267
1 12 7 SYS Backup: MML create a backup piece 29 21281
1 70 15 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch02 io done 0 21272
1 72 23 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch00 imm op 1 21258
1 131 15 SYS Backup: MML create a backup piece 4 21260
1 132 15 SYS Backup: MML write backup piece 0 21274
1 199 25 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch03 imm op 1 21279
8 rows selected.
SQL>
SQL> show parameter tape
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
backup_tape_io_slaves boolean TRUE
tape_asynch_io boolean TRUE
SQL>
3. run backup job when backup_tape_io_slaves set true
SQL> /
INST SID SERIAL# USERNAME PROGRAM MODULE MACHINE CLIENT_INFO SQL_ID B_INST B_SESS EVENT S_WAIT SPID
----- ----- ---------- ---------- ------------------------------ ------------------------------ --------------- -------------------- --------------- ------- ------- --------------------------------------------- ---------- ------------------------
1 191 3 SYS sqlplus@orcldb (TNS V1-V3) sqlplus@orcldb (TNS V1-V3) orcldb 6pduj1hhyxg0f SQL*Net message to client 0 22019
1 10 53 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch02 Backup: MML create a backup piece 16 22138
1 71 9 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch03 Backup: MML create a backup piece 15 22139
1 131 5 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch00 Backup: MML write backup piece 0 22136
1 198 5 SYS rman@orcldb (TNS V1-V3) backup incr datafile orcldb rman channel=ch01 Backup: MML create a backup piece 16 22137
SQL>
SQL> show parameter tape
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
backup_tape_io_slaves boolean FALSE
tape_asynch_io boolean TRUE
SQL>
4. conclusion.
The ‘io done’ wait event means that the database had to wait on I/O slaves to complete tape write.
Since you have the BACKUP_TAPE_IO_SLAVES parameter set to TRUE your backup is using I/O slaves.
Your backup process is waiting on those slaves to complete a request.
This should only impact your backup operation and not impact your end users, meaning unless you have a performance issue with your backup,
you can safely ignore this wait event.
最后修改时间:2022-05-11 23:30:34
「喜欢这篇文章,您的关注和赞赏是给作者最好的鼓励」
关注作者
【版权声明】本文为墨天轮用户原创内容,转载时必须标注文章的来源(墨天轮),文章链接,文章作者等基本信息,否则作者和墨天轮有权追究责任。如果您发现墨天轮中有涉嫌抄袭或者侵权的内容,欢迎发送邮件至:contact@modb.pro进行举报,并提供相关证据,一经查实,墨天轮将立刻删除相关内容。




