FAS8020 Fibre Channel 16gb throughput

[HenryUK]

Hello

i thouhgt i would try the question here before possibly raising a support ticket

as it says in the title i have a fas 8020 16GB FCwith 2x ssd/sas hybrid shelves (10.3TB)- 1 aggregate on each

also a SATA shelf with 1 aggregate (27TB) - all aggregates are around 30% full

the current disk util shows around 5% during production hours - so it is not upto much
2x cisco mds 9250i 16GBFC
HP Proliant DL 380 Gen9 with 2x 8gb fibre dual ports ports (2 ports for tape library) ( 2 ports for MDS failvoer) running veeam as out solution
4x c7000 Chassis using 8gb flex fabric

so thats pretty much the stats 

so at the moment i have Veeam looking at the current configuration of the backup jobs and the backup server they are not seeing any issues with degisn - but they are not laying blame with the Netapp - as they have glimpsed at the setup and dont see any issues with the netapp either

but here is my issue

i am getting around 150mb throughout from the Netapp to the veeam backup storage

at this time there are no overlaping jobs - backup storage is a all sas hp msa - so can easily handle this

using autosupport i am only see a 300mb peak during the backup windows - i would of thought this would be alot higher?

veeam expect to be seeing alot more than 150mb throughout which is where they are stuck - but as i have said they have not pointed the finger at anyone as we have no issues with the storage in production

could there be a setting i have missed on the MDS or the windows 2012 R2 server? ports are zoned on the netapp and mpio is set on the windows server

any ideas or tests i can make to prove the potentional read/write speed?

Cheers

[marcusgross]

Hi,

you could run a systat -x 1 during backup time to get an idea whats happening on the system.

Or schedule a perfstat and send it to the support.

How does your raid setup looks like (data/parity)?

If you want to backup your VM environment, I suggest to switch to Snapshots.

Marcus

[mikkisse]

Hi

Veeam usually displays Bottleneck for a task. What type of bottleneck you have?

What is your proxy configuration? Is your HP MSA (Backup repository) connected to the proxy?

What backup method do you use? ndb \ hotadd \ direct san? 

How much vms in your job? 

Is your backup full or incremental? ... etc

[HenryUK]

Hi All

apologies for the late reply i have been working with veeam going through all my backup locations to see what improvements i can make

but we are still on this connection to our main site

attached are the sysstat -x 2 outputs from both controllers (this is during full production) - it looks like 1 controller is doing more work but i think this is down to where the nl is connected and this is the location of our fileservers

i should say that because the hybrid shelfs are not busy i am running background compression as the storage was running at around 3-6% util all the time

it is a FAS 8020 with 2x ssd hybrid shelfs and a single nl shelf

all veeam jobs are split so i do not backup fc at the same time as nl to remove any bottle necks waiting on nl

the veeam backup runs as a storage snapshot - as i have all luns visible to the veeam proxy server

i will do a perfstat on the weekend and send this off to netapp support

i mean there are no performance problems but it seems like something is not right when it comes to backups as the speed is not a great deal different from a network (nbd) backup

tha raid setup for the hybrid shelfs are

3xSSD (CACHE DISKS) - Raid 4 - 3disks1raid

1xspare ssd

16x900gb 10k (Data Disks) - RaidDP - 16disks1raid

3x 900gb 10k (root) -  RaidDP - 3disks1raid

1x 900gb 10k spare

NL shelf

23x 1.8TB 7200k - RaidDP - 23disks2raid12disks

1x spare 1.2TB 7200k

also all fibre connections are OM4 of very short runs below 10meters so cabling looks all good

any suggestions? - before i start a netapp ticket

[marcusgross]

Hi,

you are using inline-compression on all volumes?

M,

[HenryUK]

i do not use in-line on NL slow volumes

just on the hybrid due to the low usage