Help a noob with Compression

[fry_p]

Hi All,
I saw another thread where people are getting very good speeds with the zstd compression method. Can anyone explain how I can utilize this? I am currently using the default compression method and level. Would I need to re-upload images?

[george1421]

@fry_p zstd is a newer and a bit more advanced than the gzip image compression.

Starting with FOG 1.3.5 zstd is used as the decompression engine for FOG (it can read both formats). So you will get some speed improvements over FOG 1.3.4 by just upgrading. To get the most speed out of the system you need to recapture your image using zstd. This can be as simple as deploy an image, change the image definition to zstd and then recapture the image right away. I can tell you upload is much slower than gzip, but the advantage of zstd is better data compression (on the fog server) and a much faster image deployment on dual core and better processors (on target computer).

[Wayne Workman]

@george1421 said in Help a noob with Compression:

I can tell you upload is much slower than gzip, but the advantage of zstd is better data compression (on the fog server) and a much faster image deployment on dual core and better processors (on target computer).

I had the exact opposite experience using an Optiplex 790. My upload speeds were just ridiculously fast, but deployment speeds lagged severely slower. Given my FOG Server is operating with a 100mbps connection in a VM on older equipment probably has a lot to do with it.

[fry_p]

@Wayne-Workman @george1421 That is interesting. I have model specific images, so maybe I should leave the older devices on the old compression method. I started an upload for a Dell 5440 laptop with zstd just before I left on Friday. I am eager to test the deployment on Monday!

[george1421]

@fry_p You do have to keep in mind that the target system really does all of the work for the FOG server. The fog server only manages the process and moves files about. The target does the data compression/decompression. So if you have an under powered target system the higher compression levels and algorithms may require too much processing power for the target.

For lower end systems I would even use a bit lower compression index than the fog default of 6, just to ease up on the target requirements at the expense of file size on the fog server.

[fry_p]

@george1421 Ah ok. I think I may be fine with the i5 and 8GB of ram in that model, but I do have a Optiplex 780 image that I can improve by lowering the compression then. Thanks for the tips!