ONTAP Discussions

MySQL Vm moved to netapp got really slow

joesmith

Hello Everyone,

Let me start out by saying that I am not a Database administrator. My company has been running a Dev and a Test MySQL Database on Vms running on local storage of our ESXi hosts. Each Database is made up of 3 Vms. 2 days ago I Migrated the 6 VMs from the local storage of the ESXi hosts onto a connected Datastore that is located on our ontap. Once we turned on the MySQL services the Databases were unusably slow. Our network connections we over saturated and it casued a lot of other issues.

 

My question for you all today is is there a better way to do this? 

 

We have 4 1G network cables connected to the ESXi host. Our netapp is connected to the switches with 4 10Gb cables? If I increased the network connection from the ESXi to 2 10Gb connection would that help?

 

Should the Databases not be located on VM? Can I build them directly on the netapp similar to the Microsoft SQL application?

 

I know this may require a big explanation so thank you for reading through this!

 

Thanks,

Joe

 

6 REPLIES 6

jcolonfzenpr

It's difficult without knowing your infrastructure (storage hardware? disk speed? ssd? sata? sas?) but 10g from the server instead of 1G will help it alot also if possible increasing mtu to 9k on the portgroup/vlan? used to connect to the storage (9k end to end).

Jonathan Colón | Blog | Linkedin

joesmith

We have a Fas2750 with sas drives in it. We do not have jumbo frame  set up on our network, however that is an eventual goal. I am not able to enable jumbo frame for the one connection,right?

paul_stejskal

Which protocol? Local disk will always be faster than networked disks. Have you run qos statistics volume latency show?

 

What does UM show? https://kb.netapp.com/Advice_and_Troubleshooting/Data_Infrastructure_Management/Active_IQ_Unified_Manager/How_to_monitor_volume_latency_from_ActiveIQ_...

joesmith

This volume is connected via NFS.

The Latency is currently sitting right around .3 with a spike up to 20 every once in a while. However the DB that is still on it is not actively being used at the moment. 

jcolonfzenpr

Any other workload running on the storage?

 

Maybe Cifs Shares?

 

Best,

Jonathan Colón | Blog | Linkedin

paul_stejskal
Is the VM itself over NFS or is the Linux VM using NFS? If it's NFS at the VMware level, that's going to be a bit faster. The reason for it is because metadata lookups happen in RAM vs having to send via NFS calls to the storage. https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/High_throughput_differences_between_local_storage_and_NAS_storage
Announcements
NetApp on Discord Image

We're on Discord, are you?

Live Chat, Watch Parties, and More!

Explore Banner

Meet Explore, NetApp’s digital sales platform

Engage digitally throughout the sales process, from product discovery to configuration, and handle all your post-purchase needs.

NetApp Insights to Action
I2A Banner
Public