INST_ID GROUP_NUMBER OPERA PASS STAT POWER ACTUAL SOFAR EST_WORK EST_RATE EST_MINUTES ERROR_CODE CON_ID
---------- ------------ ----- --------- ---- ---------- ---------- ---------- ---------- ---------- ----------- -------------------------------------------- ----------
1 1 REBAL COMPACT RUN 1 1 1378 0 991 0 0
1 1 REBAL REBALANCE DONE 1 1 3930 3930 0 0 0
1 1 REBAL REBUILD DONE 1 1 0 0 0 0 0
INST_ID GROUP_NUMBER OPERA PASS STAT POWER ACTUAL SOFAR EST_WORK EST_RATE EST_MINUTES ERROR_CODE CON_ID
---------- ------------ ----- --------- ---- ---------- ---------- ---------- ---------- ---------- ----------- -------------------------------------------- ----------
1 1 REBAL COMPACT REAP 1 1 1400 0 0 0 0
1 1 REBAL REBALANCE DONE 1 1 3930 3930 0 0 0
1 1 REBAL REBUILD DONE 1 1 0 0 0 0 0
As seen above the COMPACT phase run with some work and come to a conclusing (REAP status).However, the track concept is not valid for SSD (see here and here). As such the compact phase could be completely avoided to save time on rebalance. This could be done by setting "_DISABLE_REBALANCE_COMPACT=false" on ASM instance (pre-12c systems) or by setting "_rebalance_compact'='FALSE' on the ASM diskgroup (12c and up). The key differences is (beside different names) one is for ASM instance while other is for ASM diskgroup.
When this is done the rebalance operation skip the compact phase and bring it to conclusion straight away. Same operation above with _rebalance_compact set to false on the disk group shows the folllowing.
INST_ID GROUP_NUMBER OPERA PASS STAT POWER ACTUAL SOFAR EST_WORK EST_RATE EST_MINUTES ERROR_CODE CON_ID
---------- ------------ ----- --------- ---- ---------- ---------- ---------- ---------- ---------- ----------- -------------------------------------------- ----------
1 1 REBAL COMPACT WAIT 1 1 0 0 0 0 0
1 1 REBAL REBALANCE RUN 1 1 3892 3930 1857 0 0
1 1 REBAL REBUILD DONE 1 1 0 0 0 0 0
INST_ID GROUP_NUMBER OPERA PASS STAT POWER ACTUAL SOFAR EST_WORK EST_RATE EST_MINUTES ERROR_CODE CON_ID
---------- ------------ ----- --------- ---- ---------- ---------- ---------- ---------- ---------- ----------- -------------------------------------------- ----------
1 1 REBAL COMPACT REAP 1 1 0 0 0 0 0
1 1 REBAL REBALANCE DONE 1 1 3930 3930 0 0 0
1 1 REBAL REBUILD DONE 1 1 0 0 0 0 0So this looks like a best practice to have on systems using flash disks. But as this is a hidden parameter and consulting Oracle support on the merit of this didn't give confidence. Oracle support's comment was "we don't have official information from development that compact phase can be disabled as per the Best practices". There maybe some validity on this as 2198773.1 shows setting _rebalance_compact to false could cause issues with regard to rebalance operation completing (doesn't mention what type of disks were used). MOS doc 1573768.1 also advises to set _DISABLE_REBALANCE_COMPACT=TRUE. However, testing above scenarios on 19c (multuple times) didn't result in any issues.All in all there's no conclusive advise on the matter. Lot of Oracle community posts for setting _rebalance_compact to false while offical Oracle support and documents state otherwise.
Related Metalink notes
WARNING: rebalance not completed for group [ID 2198773.1]
What is ASM rebalance compact Phase and how it can be disabled [ID 1902001.1]
Can ASM “_DISABLE_REBALANCE_COMPACT=TRUE" Be Used With NetApp SAN Environment? [ID 1573768.1]
