Transcript Hardware Tuning

CS5226
Hardware Tuning
Application
Programmer
(e.g., business analyst,
Data architect)
Application
Sophisticated
Application
Programmer
Query Processor
(e.g., SAP admin)
Indexes
Storage Subsystem
Concurrency Control
Recovery
DBA,
Tuner
Operating System
Hardware
[Processor(s), Disk(s), Memory]
2
Outline

Part 1: Tuning the storage subsystem



RAID storage system
Choosing a proper RAID level
Part 2: Enhancing the hardware
configuration
3
Magnetic Disks
tracks
platter
spindle
read/write
head
actuator
disk arm
Controller
1956: IBM (RAMAC) first
disk drive
5 Mb – 0.002 Mb/in2
35000$/year
9 Kb/sec
1980: SEAGATE
first 5.25’’ disk drive
5 Mb – 1.96 Mb/in2
625 Kb/sec
1999: IBM MICRODRIVE
first 1’’ disk drive
340Mb
6.1 MB/sec
disk interface
4
Magnetic Disks

Access Time (2001)





Controller overhead (0.2
ms)
Seek Time (4 to 9 ms)
Rotational Delay (2 to 6
ms)
Read/Write Time (10 to
500 KB/ms)
Disk Interface


IDE (16 bits, Ultra DMA 25 MHz)
SCSI: width (narrow 8 bits
vs. wide 16 bits) frequency (Ultra3 - 80
MHz).
5
Storage Metrics
DRAM
Unit Capacity
Unit Price
$/Gb
Latency (sec)
2GB
1600$
800
1.E-8
Disk
18GB
467$
26
2.E-3
Tape Robot
14x70Gb
20900$
21
3.E+1
(15k RPM)
Bandwidth
1000
Kaps
Maps
Scan time
1.E+6
1.E+3
2
(Mbps)
(sec/Tb)
40
(up to
160)
470
23
450
40
(up to 100)
3.E-2
3.E-2
24500
6
Hardware Bandwidth
System Bandwidth Yesterday
in megabytes per second (not to scale!)
40
15
per disk
Slide courtesy
of J. Gray/L.Chung
The familiar
bandwidth
pyramid:
133
422
The farther
from the CPU,
the less
the bandwidth.
Hard Disk | SCSI | PCI | Memory | Processor
7
Hardware Bandwidth
System Bandwidth Today
in megabytes per second (not to scale!)
The familiar
pyramid is gone!
PCI is now the
bottleneck!
26
26
160
133
1,600
In practice,
3 disks can reach
saturation using
sequential IO
Hard Disk | SCSI | PCI | Memory | Processor
26
Slide courtesy
of J. Gray/L.Chung
8
RAID Storage System

Redundant Array of Inexpensive Disks



Combine multiple small, inexpensive disk
drives into a group to yield performance
exceeding that of one large, more
expensive drive
Appear to the computer as a single virtual
drive
Support fault-tolerance by redundantly
storing information in various ways
9
RAID 0 - Striping

No redundancy


No fault tolerance
High I/O performance

Parallel I/O
11
RAID 1 – Mirroring

Provide good fault tolerance



Works ok if one disk in a pair is down
One write = a physical write on each disk
One read = either read both or read the less busy one

Could double the read rate
12
RAID 3 - Parallel Array with
Parity



Fast read/write
All disk arms are synchronized
Speed is limited by the slowest disk
13
Parity Check - Classical




An extra bit added to a byte to detect errors
in storage or transmission
Even (odd) parity means that the parity bit is
set so that there are an even (odd) number
of one bits in the word, including the parity
bit
A single parity bit can only detect single bit
errors since if an even number of bits are
wrong then the parity bit will not change
It is not possible to tell which bit is wrong
14
RAID 5 – Parity Checking


For error detection, rather than full
redundancy
Each stripe unit has an extra parity stripe

Parity stripes are distributed
15
RAID 5 Read/Write

Read: parallel stripes read from multiple disks


Good performance
Write: 2 reads + 2 writes




Read old data stripe; read parity stripe (2 reads)
XOR old data stripe with new data stripe.
XOR result into parity stripe.
Write new data stripe and new parity stripe (2
writes).
16
RAID 10 – Striped Mirroring

RAID 10 = Striping + mirroring


A striped array of RAID 1 arrays
High performance of RAID 0, and high tolerance of RAID 1
(at the cots of doubling disks)
.. More information about RAID disks at http://www.acnc.com/04_01_05.html
17
Hardware vs. Software RAID

Software RAID




Software RAID: run on the server’s CPU
Directly dependent on server CPU performance and load
Occupies host system memory and CPU operation,
degrading server performance
Hardware RAID



Hardware RAID: run on the RAID controller’s CPU
Does not occupy any host system memory. Is not operating
system dependent
Host CPU can execute applications while the array adapter's
processor simultaneously executes array functions: true
hardware multi-tasking
18
RAID Levels - Data
Settings:
accounts( number, branchnum, balance);
create clustered index c on accounts(number);



100000 rows
Cold Buffer
Dual Xeon (550MHz,512Kb), 1Gb RAM, Internal
RAID controller from Adaptec (80Mb), 4x18Gb
drives (10000RPM), Windows 2000.
19
RAID Levels - Transactions
No Concurrent Transactions:
 Read Intensive:
select avg(balance) from accounts;

Write Intensive, e.g. typical insert:
insert into accounts values (690466,6840,2272.76);
Writes are uniformly distributed.
20
RAID Levels
Throughput (tuples/sec)
Read-Intensive

80000
60000
40000
20000
0
SoftRAID5
RAID5
RAID0
RAID10
RAID1
Single
Disk

SQL Server7 on
Windows 2000
(SoftRAID means
striping/parity at host)
Read-Intensive:

Throughput (tuples/sec)
Write-Intensive
160
120
80

40
Write-Intensive:

0
SoftRAID5
RAID5
RAID0
RAID10
RAID1
Single
Disk
Using multiple disks
(RAID0, RAID 10, RAID5)
increases throughput
significantly.
Without cache, RAID 5
suffers. With cache, it is ok.
21
Comparing RAID Levels
RAID 0
RAID 1
RAID 5
RAID 10
Read
High
2X
High
High
Write
High
1X
Medium
High
Fault
tolerance
No
Yes
Yes
Yes
Disk
utilization
High
Low
High
Low
Key
problems
Data lost
when any disk
fails
Use double the
disk space
Lower throughput
with disk failure
Very expensive, not
scalable
Key
advantages
High I/O
performance
Very high I/O
performance
A good overall
balance
High reliability with
good performance
22
Controller Pre-fetching No,
Write-back Yes

Read-ahead:




Prefetching at the disk controller level.
No information on access pattern.
Better to let database management system do it.
Write-back vs. write through:

Write back: transfer terminated as soon as data is
written to cache.


Batteries to guarantee write back in case of power failure
Write through: transfer terminated as soon as
data is written to disk.
23
SCSI Controller Cache - Data
Settings:
employees(ssnum, name, lat, long, hundreds1,
hundreds2);
create clustered index c on
employees(hundreds2);
 Employees table partitioned over two disks; Log on a
separate disk; same controller (same channel).
 200 000 rows per table
 Database buffer size limited to 400 Mb.
 Dual Xeon (550MHz,512Kb), 1Gb RAM, Internal RAID
controller from Adaptec (80Mb), 4x18Gb drives (10000RPM),
Windows 2000.
24
SCSI (not disk) Controller
Cache - Transactions
No Concurrent Transactions:
update employees set lat = long, long = lat
where hundreds2 = ?;


cache friendly: update of 20,000 rows (~90Mb)
cache unfriendly: update of 200,000 rows (~900Mb)
25
SCSI Controller Cache

2 Disks - Cache Size 80Mb

Throughput (tuples/sec)
2000
no cache
cache
1500
SQL Server 7 on Windows
2000.
Adaptec ServerRaid
controller:


1000

500

0
cache friendly (90Mb)
cache unfriendly (900Mb)
80 Mb RAM
Write-back mode
Updates
Controller cache increases
throughput whether
operation is cache friendly
or not.

Efficient replacement policy!
26
Which RAID Level to Use?

Data and Index Files



RAID 5 is best suited for read intensive apps or if
the RAID controller cache is effective enough.
RAID 10 is best suited for write intensive apps.
Log File

RAID 1 is appropriate


Fault tolerance with high write throughput. Writes are
synchronous and sequential. No benefits in striping.
Temporary Files

RAID 0 is appropriate.

No fault tolerance. High throughput.
27
What RAID Provides

Fault tolerance







It does not prevent disk drive failures
It enables real-time data recovery
High I/O performance
Mass data capacity
Configuration flexibility
Lower protected storage costs
Easy maintenance
28
Enhancing Hardware Config.

Add memory


Cheapest option to get better performance
Can be used to enlarge DB buffer pool




Better hit ratio
If used for enlarge OS buffer (as disk cache), it
benefits but to other apps as well
Add disks
Add processors
29
Add Disks

Larger disk ≠better performance


Add disks for





Bottleneck is disk bandwidth
A dedicated disk for the log
Switch RAID5 to RAID10 for update-intensive apps
Move secondary indexes to another disk for writeintensive apps
Partition read-intensive tables across many disks
Consider intelligent disk systems

Automatic replication and load balancing
30
Add Processors

Function parallelism

Use different processors for different tasks


GUI, Query Optimisation, TT&CC, different types of apps,
different users
Operation pipelines:



E.g., scan, sort, select, join…
Easy for RO apps, hard for update apps
Data partition parallelism

Partition data, thus the operation on the data
31
Parallelism

Some tasks are easier to parallelize



E.g., join phase of GRACE hash join
E.g., scan, join, sum, min
Some tasks are not so easy

E.g., sorting, avg, nested-queries
32
Summary

We have covered:

The storage subsystem


RAID: what are they and which one to use?
Memory, disks and processors

When to add what?
33
Database Tuning
Database Tuning is the activity of
making a database application run
more quickly. “More quickly” usually
means higher throughput, though it
may mean lower response time for
time-critical applications.
34
Tuning Principles





Think globally, fix locally
Partitioning breaks bottlenecks (temporal
and spatial)
Start-up costs are high; running costs are
low
Render onto server what is due onto Server
Be prepared for trade-offs (indexes and
inserts)
35
Tuning Mindset








Set reasonable performance tuning goals
Measure and document current performance
Identify current system performance
bottleneck
Identify current OS bottleneck
Tune the required components eg:
application, DB, I/O, contention, OS etc
Track and exercise change-control procedures
Measure and document current performance
Repeat step 3 through 7 until the goal is met
36
Goals Met?





Appreciation of DBMS architecture
Study the effect of various components
on the performance of the systems
Tuning principle
Troubleshooting techniques for chasing
down performance problems
Hands-on experience in Tuning
37