Latest FOG 0.33b
-
r1283 and r1284 released.
r1283 Make BootMenu slightly more dynamic. Uses, for verification, the verifyCreds() function rather than having remake it. Uses extra get variable, but all auth should go through here and just pass a new get, verify which get to send to right places.
r1284 just removed a commented out if statement in the BootMenu.class.php
-
r1285 released.
Adds quite a few components.
GUI First:
Only change here is, on the FOG Configuration Page in FOG Settings (FOG Configuration->FOG Settings) You used to have to type search or list in the field. While this worked, it left the work up to the user to actually type something properly. I think it defaulted to search, but in either case, it’s now setup as a Drop Down option. Thank you BPSTravis. He’s new to the forums, but has been hunkered in the webchat.freenode.net #fogimaging irc channel for, I guess, quite some time. So as he suggested and gave me the code base to work with, i added it with ease.init.xz:
The change here is really just to the /etc/fstab to auto mount the efivarsfs system. I don’t think it will work yet, but it is there for when we finally get uefi accessible. The other change for the init.xz is I’ve updated efibootmgr to 0.6.0.bzImage:
I read a LOT of things dealing with efi/uefi lately and I found a few issues in my bzImage. I went through and corrected where needed. Just updated so efi should, when we get there again, load/mount properly from a kernel aspect now as well.
NOTE: it’s possible already to use this bzImage if we get efi pxe loading, so no changing anything is needed. I’ve added the stub support almost for as long as I’ve been building kernels for FOG.TomElliott.config:
Updated and posted in both the src/buildroot tree and the kernel tree to reflect the messages you see in the bzImage section.snponly.efi:
I’ve built the snponly.efi ipxe (pxelinux.0 equivalent) file using the same embed script we see in use for the undionly.kpxe file.undionly.kpxe:
I’ve updated my “git” tree of ipxe and rebuilt this binary. That’s it. I need to keep updating until, at least, we finally have UEFI iPXE boot menu just like the original. This file is just the element of the current updates: 2014-03-05efibootmgr:
I’ve updated from 0.5.4 to 0.6.0. That’s it. I believe this version actually supports both sysfs_efivars and efivarsfs. This is not currently scripted anywhere, but it is there for when we are finally able to use it.src/buildroot:
All the scripts have been, properly now, updated so you all could try building your own init.xz should you so choose. The config for this is updated as well. Preferably, if you copy the folder, just copy fog.buildroot.config as .config.Hopefully this is descriptive enough.
Please feel free to chime in if you can help with ipxe uefi boot or other things needed.
Thank you,
-
r1286 released.
Fixes my dumb broken undionly.kpxe file. Sorry about that.
-
r1287 released.
Fixes an issue where the Storage node Sub menu items were not populating properly.
-
r1288 released.
Multicast Port generation, which is randomly generated as long as it doesn’t match the currently used port, has been fixed to only use ports between 49152 and 65534 (even numbered only) so we don’t run into a possible port entry on port 80 (web), 20, or other evenly needed ports.
-
r1289 released.
Re insert kernel dir for updating. If updating chown -R fog <fogweb>/service/ipxe command to allow the ftp of t
he file. Also update FOG Configuration->FOG Settings->FOG_TFTP_PXE_KERNEL_DIR field to to full path of the ipxe folder (e.g. /var/www/fog/service/ipxe, or /var/www/html/fog/service/ipxe).Sorry about the oversight.
-
does TFTP_PXE_CONFIG_DIR also need updating?
-
No, I’ve removed it completely. It’s not needed at all.
-
r1290 released.
Add build scripts. Adjust for 32/64 bit. Buildroot 32 doesn’t seem to work…yet and fails specifically on chntpw, will need to look into. Add ipxe build scripts as well. Add 32 bit kernel. Fix error thrown for hwinfo page and remove the “submenu” link as it’s unnecessary. Edit changelog. Remove .config from buildroot, must use either 32 or 64, but only 64 is known to work at the moment.
My 32 bit init.xz, currently, is failing on chntpw building. Will need to look into.
-
r1291 released.
Fixes the patch for chntpw so it works with 140201 version.
Adds the init_32.xz file.
-
r1292, 1293 released.
Fixes a couple issues found on the boot menu. Sorry I make mistakes too. The issues reported were no sysinfo menu option and snapin tasks cause boot loop. These should be corrected for now.
Thank you,
-
r1294 released.
Just a few gui tweaks here and there.
Thank you,
-
r1295 released.
Couple changes/edits/additions here.
First, Hide menu is back in the picture. If the checkbox is selected, it will hide the menu, but still show the FOG Background. It tells the user it’s booting the system. It does have a key to get to the menu if you need to. If you press the key (notice I just said one) it will prompt for login. Then, when logged in, it will give the menu.
As I’m testing with VM’s I need to use:
[code]sanboot --no-describe --drive 0x80[/code]
To boot my systems when testing. This isn’t always the case. If your systems are hanging change these lines (there’s two of them) to just say:
[code]exit[/code]
It should release back to the bios and continue booting per your bios settings. Many systems are fine without the change, but some require it.It fixes the bzImage back to 64 bit, if that was broken I don’t know.
HostManagementPage has the add functions set to allow hooking. Just as a proof of concept.
GroupManagementPage. Starting to implement checking of all hosts and reading that data. Found a simple method to doing so. It, too, is proof of concept though. Just Image and Location is checked (if you have location enabled/installed). Will work on other elements as well.
Enjoy!
Thank you,
-
r1296 released.
Adds a sequence selector based on what is currently available in ipxe. This required a schema change from 42 to 79 as we store the sequences in the Database.
The selector is on PXE Menu settings page.
Thank you very much BPSTravis for suggesting and helping with the code on this.
-
r1297 and 1298 released.
r1297 adds some hooking to the Host management page. Adds hooking to the UserManagementPage, adds hooking to the Dashboard page.
r1298 released. Some more hooking in HostManagement, but biggest change is on the FOG Settings part of the FOG Configuration page. All on/off values are now check boxes. QUICKREG_IMG_ID is a select box of all the images. OS_ID is not set as it’s set by the IMG_ID anyway, but it will display the os type of the selected image, if one is selected. DEFAULT_LOCALE is now a select box as well, and written in their respective languages.
-
r1299 released.
Fixes the FOG Configuration Page QUICKREG_IMG_ID so it actually stores the data. Fixes the green fog HH:MM section to maintain their size when selected for typing.
-
r1300 released.
Just fixes the login page language selection parts so they work appropriately again.
-
r1301 released.
Adds a lot of hooking elements to the GroupManagementPage. This is my latest project is to add hooking to all the of the pages. I’ve completed, where possible, the DashBoardPage, and added some to the HostManagementPage, but will work on one page at a time as it is rather time consuming. So GroupManagementPage isn’t 100% but I needed a break for a little bit. But basically all area’s down to alo on the service settings element of the group page are now hooking enabled. I’m sure it’s not a perfect system, but it gets the job done for now.
Updated the TomElliott.config.32/64 files to remove MTRR support. Updates the bzImage, bzImage32 files to reflect as well.
-
So… What is hooking?
I’ve seen all these updates about it…
-
this is obviously old and outdated but give you the idea:
[url]http://www.fogproject.org/wiki/index.php/Hook_System[/url]you could potentially get fog to “look” how you want it to (within reason) without hard coding your changes and making changes to the core code of fog and for example when it’s time for a fog upgrade, it would be easier to implement your “hooks” back in and in “theory” wouldn’t have to do anything to do that.
