Jump to content
  • 0

Is there any way to ensure some directories is using the SSD cache


Javen

Question

Hello

I'm new to Stablebit and may ask something very basic. My question is about best way to use SSD as cache for HDD. Here is my rig:

  • Windows 11
  • Intel P4510 SSD 2T x1
  • HC550 HDD 16T x1
  • Element HDD 14T x1

What I want to achive is that I would like to use a SSD as cache to accelerate the HDD. To achieve that purpose, I assigned all the disks into one pool and set P4510 as SSD amd others as archive in SSD optimizer plugin. (I also removed the driver letter assigned to these disks as I would like to use them only from the pool)

I store all data beside system into the pool, including movies, games, virtual machines, etc... I also set duplications for some of them.(e.g. virtual machines)

However, I found virtual machines speed is slow. e.g. starting apps in windows virtual machine may cost 10+ seconds. It's just like old days when we don't have SSD. Though I don't have evidence the virtual machine is running directly from HDD instead of SSD cache, I heavily suspect that.

 

So my questions are:

1. Is there any way to force some directory to use cache? Or if I use VM frequently enough, will diskpool copy it to cache eventually? (I guess the caching algorithm is based on the access frequence, right?)

2. If no, what I can do to achieve that goal?(Sure I understood create a pool purely based on SSD will achieve that, but let's not go to that firstly. As I still want cache enabled for HDD and the SSD slot on my mATX board is really limited, only 2.) 

Link to comment
Share on other sites

3 answers to this question

Recommended Posts

  • 0
On 1/7/2023 at 6:09 AM, Christopher (Drashna) said:

Just a heads up, the SSD Optimizer balancer is mostly meant for writes, and writes of new files.  It doesn't effect reads, unless those files are still on the SSD (or locked there). 

If you're looking to boost the read speeds, you'd want something like primocache.

Thanks for clarification.

 

So with SSD cache, new files are firstly writen to SSD but will move to HDD later by SSD Optimizer balancer? (so that read speed does not boost)

 

The thing is that I have lots Windows modern apps(mostly Xbox app downloaded games from XGPU) and WSL2 based linux distros and docker desktop. I learnt here that above usage is problematic. (However, I don't quite understand the limitation in WSL2, since after I put WSL2 vhdx files to the pool and run I face no issue till now)

 

To me, seems the only feasible way is to store those data to raw disk directly. However, that immediately means I lost the duplication feature for some of my critical data(e.g. docker containers/WSL2 distro) and caching in drivepool will not benefit for those data.

 

For the lost duplication feature, I have no idea now how to work around.

For caching, I could use primocache as well.

 

Last thing to check, since now I have 2 SSDs, if I want to boost both read and write, does combining using primocache & diskpool SSD cache make any sense? Or only using primocache is enough and not set any cache in diskpool.

Link to comment
Share on other sites

  • 0
On 1/8/2023 at 7:02 AM, Javen said:

So with SSD cache, new files are firstly writen to SSD but will move to HDD later by SSD Optimizer balancer? (so that read speed does not boost)

Correct.  The balancing system will move the files to the "archive" drives, based on the balancing settings. 

On 1/8/2023 at 7:02 AM, Javen said:

The thing is that I have lots Windows modern apps(mostly Xbox app downloaded games from XGPU) and WSL2 based linux distros and docker desktop. I learnt here that above usage is problematic. (However, I don't quite understand the limitation in WSL2, since after I put WSL2 vhdx files to the pool and run I face no issue till now)

Putting the VHDX files should be fine, since it's a simple file. 

However, WSL and the like use special drivers on top of the file system, and may do "non standard" things.  Unfortunately, a lot of these are not documented (or at least, not documend in such a way that is useful for anyone implementing a file system or emulating it).  So, a lot of reverse engineering would be needed, at best. 

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Answer this question...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...