[quote=“loosus456, post: 38867, member: 26317”]
So, here would be the requests:
[LIST]
[]A way in the FOG web console to determine which computer(s) are holding up the rest. Is there a way that FOG can determine which computer(s) are the problem?
[]Once it has been determined which computer(s) are the problem, a way to kill off just these computers from the multicast task and let the other computers continue imaging at the higher speed.
[/LIST]
Are these realistic?[/quote]

I use Iperf to determine wich computer have bad bandwith. I’ll try to added iperf on iPXE.

Regards,
Ch3i.

[quote=“Tom Elliott, post: 38874, member: 7271”]There already is that feature, under the host and under reports.[/quote]

OK… I left my brain in my bed, I think… 😄

if the only reason you want to use custom init’s is because of the xp imaging issue, that problem is fixed in the development version of fog. just use the latest dev version and your problem should go away.

redo backup / ghost / acronis trueimage etc can all do this 🙂

fog is a very good piece of software for imaging networked PCs.

slightly problematic are laptops that are normally wireless but these are things you accept.

possibly if the developers spent all their time creating what has already been created the main product (network booting and imaging) might suffer.

also… if this is windows your imaging you could use the windows tools to make an unattended USB that automatically installs windows.

I love the direction fog is going 🙂

I already saw, merged, and fixed in both git and svn.

I’ve created a table in FOG that does, more or this, exactly what’s being requested.

Because of this, for the most accurate of data, please use the proper filenames as it’s the best method I can come up with.

If you’re interested in finding out what files work and such, hence the idea behind the generation of this table, rather than setting your dhcp filename option to undionly.kpxe, start out with ipxe.pxe. Test your systems, and find out if the systems can boot under ipxe.pxe.

If they can, there’s little more you need to do.

If not all your systems work, then try the undionly.kpxe file. The table will read the new file name and create a new entry if it makes it into the menu. If it doesn’t make it into the menu it won’t enter it.

The basic construct of the table currently is:

Product, Manufacturer, Boot file, and mac address.

It will not keep making entries in the table for every mac address. It will only make new entries if the filename, product, and the manufacturer does not already exist.

This way, let’s say you changed undionly.kpxe to ipxe, you may end up with an entry for the same model of system for multiple files.

The ideology of this table is to eventually allow centralized place for fog and the ipxe team to better know what systems work and with which files.

Of course, when this gets entered, it will be completely optional to send this data to us.

Once we have the “central” table setup, the table will allow us (the developers) to hopefully code in what systems and what files to use for those systems.

For example, if we know the ipxe.pxe works for the Dell ts1700, and you’re booting to undionly.kpxe, We can supposedly cause the undionly.kpxe to rechain to ipxe.pxe. This would allow better operation of the nic during the ipxe state.

May be it is the wrong thread. “Configuration” is only a (bad) suggestion for the new feature name. By configuration I mean deployment configuration or computers configuration. It may be called “preconfigured tasks”.

We have several images for the same hosts, with different kinds of OS (linux, Windows, Windows server…) . Each time we switch from one to another (it can be twice or thrice a day) we have to do an image association (and OS before 1.2.0) and sometimes we have to do it quickly. What I mean is a superset of groups and images. By using it it can reduce time a little and error probability a lot.

As I said this feature should be useful for teaching environment.

I had a look to the code, but I am not very good at PHP and I haven’t much time for it. I would need some explanation of the code architecture to feel more confident with it.
Modifying groups implementation would be an error.
I think, it can be implemented just (always easy to say) by adding a table in database and a new page.
This page could be a minor (I am optimistic) change of the group page because all the actions would be the same.

Sorry if it is still not understandable, but I am not used to write english.

Tom is a wonderful contributor to the project (with Junkhacker from what i saw), a great support technician and very professional worker.
I have nearly the same daily work as him (i’m work in 3 college with around 500 computers in total, maybe more) and i couldn’t spend the tenth of his time on FOG (testing, debugging, improving, solving, etc.).

Everybody should understand that FOG is free and so, people working on the project doesn’t earn money for the time spend on it.
They only earn, other than the last added “Donate” advances task, is thankfulness from us (computer user, technician, administrator and company owner) and a bit of our time for testing, and bring them back the few bugs we could see.

Again, Tom, thank you. 😉

./run.sh: 2: .: Can’t open /usr/share/fog/lib/funcs.sh

Thanks Tom for answering, I am truly honored just by this.
Here is my humble response
1.[FONT=Times New Roman] [/FONT]I had the fog windows client installed on xp , and when I was scheduling task for immediate image making , it rebooted on its own to start the process ,on windows 8.1 that didn’t happen so I figured something was blocking the connection between fog server and the client, that’s why I thought about a visible “handshake " something to tell me if the connection is made or not.
2.[FONT=Times New Roman] [/FONT]Clarifying this matter is essential, because the first thing I did was to change the management password to my login password , because hey it’s the same username how come the password is different ?.. so I guess to continue my line of thought the username for internal fog staff (FTP/TFTP and others ) should be different from the default web user login , say “fogftp” or “lighthouse” or whatever… Just as long as it doesn’t confuse us innocent users…
3.[FONT=Times New Roman] [/FONT]”[FONT=Tahoma]How can we “automount” a non-existent folder?[/FONT]" that’s exactly my point , fog server should verify the existing of the path , and if something fails , it should [U]deny[/U] the change of it , Now I can click Update , and no verification takes place. Of course this is the responsibility of the user , but at least it should notify there is something wrong
4.[FONT=Times New Roman] [/FONT][FONT=Tahoma]“FOG already checks if the image exists in the case of download tasks. … of these things fail during task creation, the task will not be generated. Is this what you’re referring to?[/FONT]” No , I had to manually change the image folder permissions and mount it things I only found out after I got “access denied” error and mounting error later on. I expect fog to check these things for me . but then again perhaps it[FONT=Arial]'[/FONT]s just me.
5.[FONT=Times New Roman] [/FONT]This is my test machine, once I get it up and runging , I am planning to install it on another machine. So that can be more of a notification then a feature request per say
6.[FONT=Times New Roman] [/FONT] I meant to [U]fish out[/U] just the errors not all the logs , I am aware this is not easy , perhaps an easier thing would be a red flag inside the web interface showing something went wrong , and when clicking it a window will open with the error details.

Thank you very much for you referring to my note , I am sure if I was a Linux Guru
I wouldn’t write half of these things, but as u can imagine, I am not [FONT=Wingdings]J[/FONT]

May the force be with you , Always.
J.

[quote=“Nik777, post: 34029, member: 649”]Hi All,

Many years ago I wrote a feature for FOG 0.28 which allowed a scheduled task to cancel other tasks.
I am now evaluating FOG 1.2, and would like to know if and how I can get the change incorporated into the mainline FOG code (or as some kind of plugin).[/quote]

Interesting that you already have written such a thing. I don’t know how useful, in today’s FOG, it really would be, but that’s not to say it couldn’t be included. I don’t know what the code looked like though so I don’t know where to begin in helping you out with
this.
[quote=“Nik777, post: 34029, member: 649”]
The problem the feature solves is that if I scheduled a virus scan at 1am, but some machines fail to wake-up-on-LAN, then when the users boot the machine later in the morning, the machine proceeds with the virus scan then, with no way of avoiding it, save going into FOG and cancelling the task.[/quote]

While the feature solves this problem, one of the Features I’ve added to FOG is to check all hosts that currently have a task. If they have not checked in at all, they will be resent the WOL packet. The reason for this? Because I’ve seen some networks will, seemingly arbitrarily, block some systems from receiving the packet. Waiting a period of time and resending the packet seems to work. That’s not to say this wouldn’t still have it’s uses, but just saying that we’ve added checks and balances to try helping with problems such as these.

[quote=“Nik777, post: 34029, member: 649”]
So I wrote a cancellation task, which I schedule for 3am. Then, any Virus scan task that has not run by 3am is automatically cancelled, so the users could use the machine when they come in.

While I only implemented the UI to support a canccellation task for the virus scan, the cancellation task can be applied to any other type of scheduled task.

Any and all comments are welcome.

Cheers!
Nik[/quote]

I think the best approach for this would to create a TaskType that specifically does this. The weird thing about this method, and I’m guessing you figured out the kinks on this, would be that Tasks are typically associated with hosts. How would one create a task that really isn’t host associated? Furthermore, how could we get a “cancellation task” so it doesn’t require you going to the Host page to schedule?

OK, so here is what’s needed to switch FOG 1.2.0 between networks (interfaces):

FOG System Settings -> FOG Configuration:
* General Settings: FOG_WOL_HOST (change ip); FOG_WOL_INTERFACE (change iface)
* Multicast Settings: FOG_UDPCAST_INTERFACE (change iface)
* NFS Server: FOG_NFS_ETH_MONITOR (change iface)
* TFTP Server: FOG_TFTP_HOST (change ip)
* Web Server: FOG_WEB_HOST (change ip) Storage Management -> DefaultMember:
* IP Address (change ip)
* Interface: (no changes) /tftpboot/default.ipxe: change ip in last line change default gateway using ip command (to make wake on lan work - it uses default gw for broadcasts):
* ip r del default - delete current
* ip r add default via <gw for second network>

Ahh right, if it resets the OU settings its best kept as it is, as each classroom has its own OU to manage programs that we push to specialised classrooms.

If there was an option to just refresh the username and password details to all hosts without modifying the OU’s that would be great.

Only reason I asked for the feature was because I had to change the settings as I updated to v1.2.0 and got caught out by the hosts storing credentials, maybe the hosts should just store the OU by default but if a different account needs to be used to join the machine to the domain then the option remains and could be activated per host via an option box…

Maybe if the username & password per host is left blank then it would use the settings from the configuration area, but i was trying to shorten my task of updating the host’s stored credentials by asking about a tick box that could enable the already specified username & password to be used.

Just a minor ask but in my opinion it would enhance FOG by making the management of it easier if the main domain adding account needed the password resetting or the account changes altogether.

Cheers

Thanks Tom S

Thanks Tom,

That should be more than enough to reduce human errors and save some precious time to all of us.

🙂

Rebuilding web side of services to be a cross-plataform is a great idea.
About the time-frame, this should not be a problem.
““this will be done when it is done”” 🙂
Thanks

it should be mentioned that in the latest fog kernels pauses have been added that require interaction when in debug mode. this allows you to read the messages that come up during the process.