Proper procedure for creating L2 cache on existing HA-Pair

Software-based VM-centric and flash-friendly VM storage + free version

Moderators: anton (staff), art (staff), Max (staff), Anatoly (staff)

Post Reply
dmarcus
Posts: 3
Joined: Fri Feb 08, 2019 12:03 am

Wed Feb 20, 2019 9:26 pm

Hi Everyone,
Looking for the proper procedure to create the L2 Cache image file from the free version (ie: using powershell scripts where necessary).

This guide seems to be pretty thorough, but it looks like the images are created in the console in the guide: https://knowledgebase.starwindsoftware. ... dance/661/
Boris (staff)
Staff
Posts: 805
Joined: Fri Jul 28, 2017 8:18 am

Thu Feb 21, 2019 6:44 pm

For creating a disk to further be used as L2 cache device, use the CreateImageFile.ps1 sample script and adjust the values as required in terms of file location, names etc.
Be sure to comment out the following lines:

Code: Select all

#$target = New-Target -server $server -alias "targetimg1" -devices $device.Name
#$target
Feel free to report your results to the community.
Sajmon
Posts: 11
Joined: Fri Mar 22, 2019 3:39 pm

Thu Mar 28, 2019 10:04 am

Actually I find that L2 cache guide confusing. What header should I edit for standalone device ? From the guide it looks like I should edit L2 cache header. Step 1 tells to create L2 device and step 3 tells to "Open the .swdsk file of the newly created device". I guess that means L2 cache device as it was created right before step 3. On the other hand I think myself, why would cache device refer to itself ? Luckily for me I have added L1 cache when creating HA device and thus I can see where is the cache referred. Another thing is that section "Adding cache to both HA devices" tells to repeat "Adding cache to standalone device" for both nodes. What is the section "Adding cache to the partner HA device" for ? Should I setup cache for standalone device and then follow "Adding cache to the partner HA device" or repeat the standalone procedure ? I understand that this is a free product and therefore it is best effort based, so I am not complaining. I am just trying to understand how it works, so I can configure this properly.
Boris (staff)
Staff
Posts: 805
Joined: Fri Jul 28, 2017 8:18 am

Thu Mar 28, 2019 2:29 pm

Step 3 implies opening the header file of the newly created L2 cache device. You open it to copy the config section that you are going to further modify and insert into your main device's config.

If you add cache to both HA disk partners, be sure to follow the part of the guide describing the action for a standalone device.
Sajmon
Posts: 11
Joined: Fri Mar 22, 2019 3:39 pm

Mon Apr 01, 2019 9:13 am

Thanks for clearing this up. I managed to setup the cache and it appears in the device properties. However I have one more question that is partly related to the cache and partly to the main volume. When I setup 2 CSV's and assign their primary owner to different node, does that split the storage processing load to both nodes or it is the same as if I had just one CSV ? Also if I have 2 CSV's, do I have to create create two L2 cache images or I can assign one to both CSV's ? Might be stupid questions I know, but I rather sound stupid now than be sorry later when it goes into production.
Boris (staff)
Staff
Posts: 805
Joined: Fri Jul 28, 2017 8:18 am

Mon Apr 01, 2019 3:10 pm

If different CSVs are used by different nodes in the cluster, the storage processing load is split between the nodes.
If you have two main devices created in StarWind VSAN, you cannot assign one single L2 cache device to both. Instead, you would need to create two separate L2 cache devices, and each of the devices gets its own L2 cache device.
Sajmon
Posts: 11
Joined: Fri Mar 22, 2019 3:39 pm

Tue Apr 02, 2019 6:56 am

That's what I thought, but thanks for clearing this up.
Boris (staff)
Staff
Posts: 805
Joined: Fri Jul 28, 2017 8:18 am

Tue Apr 02, 2019 2:01 pm

You're welcome.
You know where to find us in case of any questions. :)
Post Reply