Topics created by ChrisG

you need to speak to microleaks, blackout or Chuck Syperski,

also how about getting it on to [url]http://www.turnkeylinux.org/[/url] this should help spread FOG even further they are also tied in with [url]http://www.proxmox.com/products/proxmox-ve[/url]

Theoretically, maybe. But you’ve got to give a % of that 1GB to management and overhead. Same with the disk systems. If you are servicing 1 client, and the file is completely defragmented on the source drive and the system never has to read anything but a sequential sector on the disk, and, and, and…

If you image (unicast, not multicast) more than 1 client, you are stressing the disk system on the server and the network switch and you’re speeds will go down.

I can image about 14 Fujitsu laptops on a gigabit switch, each getting about 1.3GB per minute per client according to the clients. I can image about 30 acer tablets that are 100Mbps before I get below 1GB per minute per client.

I can do this because the disk system on the server is six 10K rpm disks in RAID 5 mode optimized for read throughput. My first test server was a desktop machine with a single 3GB sata drive, and I got 4GB per minute with 1 client, but it dropped to 1GB per minute by adding the 2nd client. The disk couldn’t handle the read requests from the 2nd client as well as a real server that has a raid controller with built in cache and multiple disks to read from.

Over the next few months I’m going to try getting my C# hat sharpened and fitted so hopefully I can help more with this.

If you have a C# developer, you can probably create a module that does this, but I don’t believe it’s built into FOG right now. Check the feature request forums and see if it’s already been suggested and do so if it has not been.

Unless you’re in a very demanding environment I feel like it would be wasted. If it’s just a matter of load balancing then I’d suggest setting up storage nodes. There are very few circumstances in which I would expect 10Gb to be the only option.

In the follow scenario your bottleneck is probably the HDD:
[COLOR=#999999]HDD[/COLOR] [COLOR=#ff0000]–SATA3–[/COLOR] [B]Server[/B] [COLOR=#339966]–1Gb–[/COLOR] [B]Swtich[/B] [COLOR=#339966]–1Gb–[/COLOR] [B]Client[/B] [COLOR=#ff0000]–SATA3–[/COLOR] [COLOR=#999999]HDD[/COLOR]

In the follow scenario your bottleneck is probably the 1Gb:
[COLOR=#800080]SSD[/COLOR] [COLOR=#ff0000]–SATA3–[/COLOR] [B]Server[/B] [COLOR=#339966]–1Gb–[/COLOR] [B]Swtich[/B] [COLOR=#339966]–1Gb–[/COLOR] [B]Client[/B] [COLOR=#ff0000]–SATA3–[/COLOR] [COLOR=#800080]SSD[/COLOR]

In the follow scenario your bottleneck is probably still the 1Gb:
[COLOR=#800080]SSD[/COLOR] [COLOR=#ff0000]–SATA3–[/COLOR] [B]Server[/B] [COLOR=#0000ff]–10Gb–[/COLOR] [B]Swtich[/B] [COLOR=#339966]–1Gb–[/COLOR] [B]Client[/B] [COLOR=#ff0000]–SATA3–[/COLOR] [COLOR=#800080]SSD[/COLOR]

In the follow scenario your bottleneck is probably the SSD, although you might have reached the 6Gb limit for SATA3:
[COLOR=#800080]SSD[/COLOR] [COLOR=#ff0000]–SATA3–[/COLOR] [B]Server[/B] [COLOR=#0000ff]–10Gb–[/COLOR] [B]Swtich[/B] [COLOR=#0000ff]–10Gb–[/COLOR] [B]Client[/B] [COLOR=#ff0000]–SATA3–[/COLOR] [COLOR=#800080]SSD[/COLOR]

Without upgrading the network connections all the way to the client, the only performance increase you would see from using 10Gb on the server side is from concurrent imaging tasks, which you could easily get by using storage nodes to load balance the tasks.

[quote=“Tim Jackson, post: 11199, member: 3381”]How did you get on with this? I have my Macs able to Netboot from my FOG server but need to get a method to apply images from inside the Netboot image…
cheers
timjak[/quote]

Well, I didn’t… I wrote a script to do it in Ubuntu, and will wait for partclone support in FOG… I don’t know where I put the script, but it was mainly backuping/restoring the GPT table (using gdisk, iirc), and then using the partclone hfs+ backup/restore utility for the partitions…

(Sorry, I don’t check the forums often enough… :D)

FOG currently supports 1 default AD to join by default, and that information can be auto populated when you specify a host should auto join the domain after imaging.

You can however put the computers into a FOG group for each sub domain, and then use the group settings to specify the domain information to use when joined the subdomain.

I think the process would be to get them all registered with FOG, put them into a group, and then specify the AD option for the group. Those hosts should auto join the sub-domain you specified rather than the default domain.

I know this isn’t exactly what you’re asking for, but it’s how you can do it with the current version of FOG (0.32).

We don’t have a release date for FOG 0.33 at this point. FOG 0.33 isn’t just a collection of bug fixes or feature updates, it’s a complete rewrite of the underlying code structure, so it’s not like we can say “feature X will have to wait until next time”. The standard response is “it’ll be released when it’s ready”.