Storage Administration - Oracle Exadata Recipes: A Problem-Solution Approach - page 262

cm01cel01: GridDisk SDATA_CD_00_cm01cel01 successfully created
cm01cel01: GridDisk SDATA_CD_01_cm01cel01 successfully created
cm01cel01: GridDisk SDATA_CD_02_cm01cel01 successfully created
... output omitted
[oracle@cm01dbm01 ~]$ dcli -g ./cell_group cellcli -e create griddisk all \
> harddisk prefix='RECO', size=67G
... output omitted
[oracle@cm01dbm01 ~]$ dcli -g ./cell_group cellcli -e create griddisk all \
> harddisk prefix='SRECO', size=36.6875G
... output omitted
Finally, create your DBFS_DG grid disks:
[oracle@cm01dbm01 ~]$ dcli -g ./cell_group cellcli -e create griddisk all \
> harddisk prefix='DBFS_DG'
cm01cel01: Cell disks were skipped because they had no freespace for grid disks:
CD_00_cm01cel01, CD_01_cm01cel01.
cm01cel01: GridDisk DBFS_DG_CD_02_cm01cel01 successfully created
cm01cel01: GridDisk DBFS_DG_CD_03_cm01cel01 successfully created
cm01cel01: GridDisk DBFS_DG_CD_04_cm01cel01 successfully created
Notice in this output that two cell disks were skipped on each storage server due to lack of free space; this is
because the first two disks in each storage server reserve 29 GB of space in each of the first two SAS disks for Exadata's
System Area. This is discussed in Recipe 1-4 and elsewhere in this text.
When complete, use cellcli to display grid disk details for one of your cell disks on one storage cell:
[oracle@cm01dbm01 ~]$ dcli -c cm01cel01 cellcli -e list griddisk where celldisk=CD_11_cm01cel01 \
> attributes name,size,offset
cm01cel01: DATA_CD_11_cm01cel01 284G 32M
cm01cel01: DBFS_DG_CD_11_cm01cel01 29.125G 528.734375G
cm01cel01: RECO_CD_11_cm01cel01 67G 425.046875G
cm01cel01: SDATA_CD_11_cm01cel01 141G 284.046875G
cm01cel01: SRECO_CD_11_cm01cel01 36.6875G 492.046875G
[oracle@cm01dbm01 ~]$
This output shows that we have five grid disks on this storage cell, each with the specified size and the byte offset
corresponding to the order in which we created the grid disks. To confirm that your grid disks are evenly distributed
and sized across all Exadata storage cells, run the Perl script in Listing 9-14.
Listing 9-14. lst09-14-griddisks.pl
#!/usr/bin/perl
open(F,"dcli -g ./cell_group cellcli -e list griddisk attributes name,size,offset|");
while (<F>) {
$tgd++;
($cd,$gd,$sz,$offset)=split(' ',$_);
$gd=substr $gd,0,index($gd,"_");
$cell{$cd}++;
$cell{$gd}++;
push @CELLS, $cd ;