NetApp Administration and Best Practice, Brendon Higgins, Proact UK

Post on 28-Nov-2014

768 views 0 download

description

Introduction to NetApp storage and tips for successful administration based on 6 years experience

Transcript of NetApp Administration and Best Practice, Brendon Higgins, Proact UK

1

2

https://communities.netapp.com/people/BrendonHiggins/blog

3

4

5

6

7

Throw it at the wall and see what sticks

8

Spend more time at the start on the storage basics or move quickly

technical detail and best practice advise.

9

10

How to share, shared storage

11

One good reason, given the presentation is for the virtual machine user

group (VMUG).

12

May still come across DoT 6 systems but they are becoming rare. Lots

of 7 mode systems out there still however. New features are only

being added to the DoT 8 cluster mode code.

13

https://communities.netapp.com/people/BrendonHiggins/blog/2013/01/

31/develop-an-understanding-of-das-nas-and-san

DAS and SAN look the same to the application and are block based

SAN means LUNs

NAS is file based and requires network protocols, NetApp supports

CIFS, NFS, FTP & HTTP

14

Key NetApp advantage

15

Still engineers do not know… Lots of wasted effort by the filer. NetApp

now includes tools to detect LUN alignment issues. Great jumping off

place to start learning more,

https://communities.netapp.com/message/58686.

16

Not such a problem in newer operating systems such as Windows

2012

17

CIFS share set up correctly but not working? See this frequently after

enabling NFS for VMware.

18

Add more ‘power’ to a system by swapping the filer heads. The data

stays on the disk and is unchanged. Simple process.

19

Scale out cluster limits change with every release, so check. SAN is

held back by Asymmetric Logical Unit Assignment (ALUA) technology.

20

Lots of wizards to guide and help with tasks. Hiding the complexity

whenever possible.

21

https://communities.netapp.com/people/BrendonHiggins/blog/2013/01/

07/introduce-netapp-interfaces--filerview-cli-system-manager-rsh-

powershell-etc

Console is the most powerful tool for daily administration.

All console sessions should be recorded

Filerview has been discontinued

System Manager is the replacement but be sure to keep it upgraded

22

23

24

25

26

Based on http://oracle-study-notes.blogspot.co.uk/2011/02/simple-

concept-map-for-some-basic.html

27

Key NetApp advantage

29

http://msdn.microsoft.com/en-

us/library/windows/desktop/bb968832(v=vs.85).aspx

New SQL DBA’s always say to me that the last backup product they

used is ‘so’ much better than this NetApp mumbo, jumbo but they fail to

understand that the underling technology is the same. In order to get

the backup to run reliable, just sort out the application environment out

and everything works just fine…

30

31

32

33

One way process…

34

Based on snapshots, easily reversible and a key NetApp feature

35

36

Sometimes ‘Sorry boss’ is not enough.

37

So many consoles…

38

39

40

The default reported state HAS to be healthy or people use accept it is

constantly ‘broken’. Check at least twice a day and resolve any issues

ASAP.

41

42

43

Day long workshop in its own right

44

45

Day long workshop in its own right

46

http://www.purplemath.com/modules/meanmode.htm

Note the mean number is not in the list of actual numbers in the range.

47

http://en.wikipedia.org/wiki/Normal_distribution

When putting forward calculated numbers as ‘actual’ results be sure to

consider what the range of data looks like and how far your number

deviates from the centre.

48

49

Day long workshop in its own right

50

An Exchange 2003 datastore on a single LUN. The aggregate had 60

300gb 15k FC disks and in Aug 2010 the FAS3070 heads were

upgraded to FAS3170. In Nov 2010, the flash cache was enabled.

Notice the average trend before and after flash cache. Also this

diagram enabled the upgrade to fly through the Cap-Ex process.

51

52

Export the disk performance to Excel. Notice that dparity & parity disks

have negligible IOPs and skew the results of any calculations so

remove them.

53

As the disks of the aggregate become more busy, the latency will

increase.

54

As the disks of the aggregate become more busy, the latency will

increase.

55

Let the tools do the heavy lifting

56

57

58

Always test for yourself what is being reported by using a known

source.

59

60

61

Create your own support issues. Get the a support number and start

the clock on the SLA timer. Call the support number after about 30

minutes, if you have not received an email or phone call. If you find

you are not getting the support you need, ask for escalation. In the

unlikely event this does not happen, ask to speak to the duty manager.

62

Learn this tool! - http://support.netapp.com/NOW/asuphome/

63

64

65

Recertification every 2 years

66

It is how I did it

67

68

69