This content has been marked as final. Show 6 replies
In an ASM diskgroup data is striped over the disks with the AU given (default 1MB).
So if you have 4 disks and a tablespace/datafile of 4 MB, the tablespace/datafile will consist of 4 extents, respectively 4 AU. One on each disk.
If your datafile is 4000MB it will consist of 4000 extents (= 4000 AUs) striped over all for disks. So each disk will hold 1000 AUs of the datafile.
This is true up to 20000 extents (respectively 20000 MB).
Now the next extent (20001 MB) for this datafile ASM is allocation will not be a single extent of 1 AU, but ASM will directly allocate 8 extents of the size of 4 AUs.
So for the datafile on disk ASM will have 20008 extents (5002 extents per disk), but since the last 8 extents are 4 AUs, the AU reservered are 20032 AUs or 5008 AUs per disks.
One extent is per disk.
it shows you have given scenario for external redundancy. If taking into account for normal redundancy,if i have 2 failgroups and 2 disk in each failgroups then each disk will contain 2 AUs. That's right?
But now after 2000 extents , 8 extents are allocated at a time. That would means each disk would contain 2 extents among 8 extents. if one extent is allocated per disk then how does striping takes place. First 4AUs will get wriiten in first extent, second 4AU will written in second extent and so on... That's right?