Asm storage doubts

Question

hi all,
In our test env we have two database in a node with only one disk group which stores around 1.5 TB of data files.
* Is that a good practice to store two DB data files in the same disk group.
* If it is in same disk group what happens when we restore a db for cloning.
a. Is that asm scans all the disks to find a free space to store the data during restore.( I do see it takes 15 hours for restoring a 800GB of a single DB data files with 6 channels specified in restore command)
*During mount stage it takes long time to open the database. Is that asm scans the entire diskgroup to match the physical structure of the files with control file.
*During restore or opening up the database there is absolutely no load on the server.
SQL> select GROUP_NUMBER,DISK_NUMBER,TOTAL_MB,FREE_MB,READS,WRITES,READ_TIME,WRITE_TIME from V$ASM_DISK;
GROUP_NUMBER DISK_NUMBER   TOTAL_MB    FREE_MB      READS     WRITES  READ_TIME WRITE_TIME
           2           5     153593       6238   61237253    3255630 1606217.37 317411.015
           2           6     153593       6248   62021450    3113196  1615818.5 300858.424
           2           7     153593       6234   60130864    3145717 1602473.04 309295.707
           2           8     153593       6253   63976632    3504263 1612983.03 325407.467
           2           0     153597       6254   61040136    3484440 1618972.86 287943.441
           2           1     153597       6268   64525211    3182327 1644359.02 295506.291
           2           2     153597       6238   64116126    4290514 1604709.78 372315.757
           2           3     153597       6277   65550559    3145245 1633260.86 289411.098
           2           4     153593       6246   62872224    3744320 1639207.33 313410.323
           2           9     153593       6248   62781328    3077681  1636946.1 295827.542Please advice on the above.
thanks,
baskar.l

Answer 1

*Is that a good practice to store two DB data files in the same disk group.Yes it is. In fact this is recommended by oracle. Oracles recommendation goes towards TWO disk groups: one for ALL database and one used as flash (or fast) recovery area.
* If it is in same disk group what happens when we restore a db for cloning.Your database will be restored there. Space is managed as a bitmap. Of course ASM has to look for free space to store the data.
. Is that asm scans all the disks to find a free space to store the data during restore.( I do see it takes 15 hours for restoring a 800GB of a single DB data files with 6 channels specified in restore command)As mentioned above: Space is managed in a bitmap structure, thus highly compact and fast.
Your restore time will be entirely dominated by getting the data from tape and save it to the disks. "Finding" free space will be less than 1% of that time.
*During mount stage it takes long time to open the database. Is that asm scans the entire diskgroup to match the physical structure of the files with control file.Either you are starting a database which was not shut down in a clean way (e.g. "shutdown abort" used to shutdown the database) or you have a lot of files. Or the storage system is slow or overloaded and so on. But asm does not scan the entire diskgroup in any way.
*During restore or opening up the database there is absolutely no load on the server.What about your storage subsystem? Did you check load there? Are the disks shared between LUNs?
Ronny Egner
My Blog: http://blog.ronnyegner-consulting.de
Edited by: Ronny Egner on Mar 31, 2010 12:12 PM