Cache Redis on Azure

-1

I contracted the Azure Cache Redis cache service, but to import the current database I had to migrate it to Premium P1 , after import I tried to go back to C0 , impossible Azure Cache Redis only makes upscale never downscale .

I need to create my database, but not C3 of the timeout. By default I believe it would run in B2 , but the database creation process is cumbersome and complex. So how do you manage to import or generate the initial database if scaling is only up?

If I do, 2 connections, one with my current redis server and another in azure, and make a select and then a set in azure, will it be fast? There are 90mil keys to generate, each one very big.

Information about Redis Azure plans link

Request to be able to downscale in the Redis service in Azure issue link

    
asked by anonymous 06.02.2017 / 15:16

2 answers

1
Unfortunately, the Azure Cache Redis (Paas) service still has some limitations like:

  • Import premium layer only;
  • Impossible to downgrade the layer, after placing in a Premium for import, will always be stuck in this.

My solution was to hire a VM with windows and install Redis Cache (it could be a linux VM as well) so the cost was lower and it performed better than Premium P1.

    
23.02.2017 / 13:09
-1

I imagine you can not do downscaling because after loading 90K keys, only a few plans will support storing all that data.

I do not know the size of your data mass, but, for example, C0 supports only 250MB of cache - this size is directly proportional to your data mass, but the ratio is not 1 for 1. p>

About timeout , if you can share the snippet of your code that is sweating to make the move, I can try to help you identify improvement points, but in any case, I recommend following the following steps:

  • Create your Azure Redis Cache service in a plan that is interesting to your business - in your case, probably in B2 ;
  • Provision a VM in the same DataCenter - if possible in the same Resource Group - of your cache service;
  • Copy your backup to this VM and import the data from there.
  • After the import is finished, delete the VM.
  • This will cause you to perform the import via "local network", making the change less timeout or loss of packages.

        
    07.02.2017 / 08:52