I/O is input /output operation to read/write data stored on disk. This operation to find data on disk can be in sequential or random fashion.
An important measure for random I/O is IOPS and important measure for sequential I/O is throughput.
Disk's response time : 1/ IOPS : time it takes to service an I/O request when the request is random and small.
IOPS : No. of IOPS per second . Affected by block size.Manufacturers will generally quote a best case IOPS (512 bytes).
Quoted values are always going to be rather optimistic . Both read and write IOPS should be measured.
To improve IOPS , disk are stacked together in an array connected to a special disk controller called a RAID controller.
RAID arrays can be built to withstand a number of drive failures.
If storage solution consist of 25 SAS drives with 200 IOPS for each SAS drive, array will be quoted as 5000 IOPS.
RAID 0 : Provide data striping across disks but no data redundancy
RAID 1 : No data striping but data redundancy of one.
RAID 5 : Provide data striping conatining a parity block.This parity block helps in redundancy.
RAID 10 : Data stripped across disks and each has a mirrored partner to provide redundancy
Disk's transfer rate : dominated by the drive's RPM and number of sectors-per-track
First check how much IOPS the applications need. Datafile and logfile will need high percentage in write IPOS and temporary will require 50/50 for read and write.
break on report
compute sum of Value on report
select METRIC_NAME,avg(AVERAGE) as "Value"
from dba_hist_sysmetric_summary
where METRIC_NAME in ('Physical Read Total IO Requests Per Sec','Physical Write Total IO Requests Per Sec')
group by METRIC_NAME;
select count (*), sum(time_waited), event from v$active_session_history group by event order by 2 desc ;
http://www.symantec.com/connect/articles/getting-hang-iops-v13
An important measure for random I/O is IOPS and important measure for sequential I/O is throughput.
Disk's response time : 1/ IOPS : time it takes to service an I/O request when the request is random and small.
IOPS : No. of IOPS per second . Affected by block size.Manufacturers will generally quote a best case IOPS (512 bytes).
Quoted values are always going to be rather optimistic . Both read and write IOPS should be measured.
To improve IOPS , disk are stacked together in an array connected to a special disk controller called a RAID controller.
RAID arrays can be built to withstand a number of drive failures.
If storage solution consist of 25 SAS drives with 200 IOPS for each SAS drive, array will be quoted as 5000 IOPS.
RAID 0 : Provide data striping across disks but no data redundancy
RAID 1 : No data striping but data redundancy of one.
RAID 5 : Provide data striping conatining a parity block.This parity block helps in redundancy.
RAID 10 : Data stripped across disks and each has a mirrored partner to provide redundancy
Disk's transfer rate : dominated by the drive's RPM and number of sectors-per-track
First check how much IOPS the applications need. Datafile and logfile will need high percentage in write IPOS and temporary will require 50/50 for read and write.
break on report
compute sum of Value on report
select METRIC_NAME,avg(AVERAGE) as "Value"
from dba_hist_sysmetric_summary
where METRIC_NAME in ('Physical Read Total IO Requests Per Sec','Physical Write Total IO Requests Per Sec')
group by METRIC_NAME;
select count (*), sum(time_waited), event from v$active_session_history group by event order by 2 desc ;
http://www.symantec.com/connect/articles/getting-hang-iops-v13
