Latest FOG 0.33b

[JiC]

[quote=“Tom Elliott, post: 24291, member: 7271”]
FOG_TFTP_PXE_CONFIG_DIR is no longer really required as we’ve moved the pxelinux.cfg directory to /var/www/fog/service/ipxe. The “config” is done dynamically from the boot.php/BootMenu.class.php files. Tasks aren’t “saved” anywhere anymore so we don’t need to ftp the 01-XX-XX-XX-XX-XX-XX files anymore.
FOG_TFTP_PXE_KERNEL_DIR should be set based on the variables during install time. If it’s not set, I don’t know why. However, this should be set to /var/www/fog/service/ipxe
FOG_TFTP_PXE_KERNEL should only contain “bzImage” as iPXE uses relative links. So when it chains to iPXE, it’s chaining to [url]http://{fogserverip}/service/ipxe/boot.php[/url]

The relative location in full form when loading bzImage and init.xz would then be:
[url]http://{fogserverip}/service/ipxe/{bzImage,init.xz}[/url]

The location of the init.xz and bzImage files is in: /var/www/fog/service/ipxe which is why you set the KERNEL_DIR parameter to that folder. If for some reason permissions weren’t setup properly, this folder should be chown -R fog.root /var/www/fog/service/ipxe
[/quote]

Thank you for all these explanations, you made many interesting changes since 0.32… I better understand how it all fits.

I only changed my DHCP server to use “undionly.kpxe” : when I started my test computer, it started with the fog kernel. It didn’t clone this computer (maybe because it’s an “Advanced Format Drive”), but cloning is ok with classic disks.

Many thanks

[warp]

I do not know if I speak well, this is what I want:

[U]Here’s what I have:[/U]

• Storage Group = default,
  in this Group -> Storage Node = DefaultMember
  image path = /images
  IP = 172.16.8.44

(This group does me nothing)

• Storage Group = NAS
  in this Group -> Storage Node = Storage_NAS
  image path = /volume1/images
  IP = 172.16.8.43

(this is where I want to save, it’s NAS Synology)

How to make:

1. The Scheduler works (because it does not work when it is Storage_NAS in master)
2. If I do not put in the master Storage_NAS but DefaultMember the Scheduler works, but I is not the path / volume1/images in the configuration file.

[warp]

[quote=“Tom Elliott, post: 24314, member: 7271”]r1343 released.

Fixes path display on Image Management Page, if you have multiple master nodes. It would only display the first node it found which would typically be the first ID in the sequence. It now pulls the master node based on the Group ID.[/quote]
If i ave multiple master nodes the service sheduler log : I don’t appear to be the group manager, I will check back later.

[Tom Elliott]

It’s not multiple master nodes. It multiple groups.

If you move the Storage_NAS into the Default group (as you’re using the server off of the default group) and make the Storage_NAS as master, all should work.

[warp]

[quote=“Tom Elliott, post: 24333, member: 7271”]It’s not multiple master nodes. It multiple groups.

If you move the Storage_NAS into the Default group (as you’re using the server off of the default group) and make the Storage_NAS as master, all should work.[/quote]
I’ve moved the Storage_NAS into the Default group, but now when i upload a image of the host it doesn’t write into the NAS under folder /volume1/images but it’s write in the first time to the DefaultMember in folder /images.

When i’ve a other group, it’s write directly to the Storage_NAS under folder /volume1/images

[Tom Elliott]

I’m still confused. Is Storage_NAS the master node in the Default group?

[Tom Elliott]

r1347 released.

More changes, namely to the BootMenu.class.php. We’ve added some elements to make it that much more similar to the “old” pxe system. Hidden menu displays the key sequence, once the key sequence is hit, it presents a login screen. Debug mode is now only accessible if HiddenMenu is enabled AND the keysequence/auth scheme are met. This is because giving, basically, anybody inherit access to debug can be pretty dangerous.

There were a few more revision changes that basically implemented this and fixed a few other areas on the BootMenu.class.php file. Just giving you all a heads up.

[Fernando Gietz]

When I do a fresh install under Centos and I update the DB from the browser, I have problems with the SQLs. It’s easy solve it:

[CODE]138c138
< “INSERT INTO " . DATABASE_NAME . ".users VALUES (‘’,‘fog’, MD5(‘password’),‘0000-00-00 00:00:00’,‘’)”,

“INSERT INTO " . DATABASE_NAME . ".users (uName,uPass, uCreateDate, uCreateBy) VALUES (‘fog’, MD5(‘password’),‘0000-00-00 00:00:00’,‘’)”,
140c140
< “INSERT INTO " . DATABASE_NAME . ".supportedOS VALUES (‘’,'”._(“Windows XP”).“', ‘1’)”,

---

“INSERT INTO " . DATABASE_NAME . ".supportedOS (osName,osValue) VALUES ('”._(“Windows XP”).“', ‘1’)”,
142c142
< “INSERT INTO " . DATABASE_NAME . ".schemaVersion VALUES (‘’,‘1’)”

---

“INSERT INTO " . DATABASE_NAME . ".schemaVersion (vValue) VALUES (‘1’)”
147c147
< “INSERT INTO " . DATABASE_NAME . ".supportedOS VALUES (‘’,'”._(“Windows Vista”).“', ‘2’)”,

---

“INSERT INTO " . DATABASE_NAME . ".supportedOS (osName,osValue) VALUES ('”._(“Windows Vista”).“', ‘2’)”,
257c257
< “INSERT INTO " . DATABASE_NAME . ".supportedOS VALUES (‘’,‘Other’, ‘99’)”,

---

“INSERT INTO " . DATABASE_NAME . ".supportedOS (osName,osValue) VALUES (‘Other’, ‘99’)”,
844c844
< (ngmMemberName, ngmMemberDescription, ngmIsMasterNode, ngmGroupID, ngmRootPath, ngmIsEnabled, ngmHostname, ngmMaxClients, ngmUser, ngmPass )

---

           (ngmMemberName, ngmMemberDescription, ngmIsMasterNode, ngmGroupID, ngmRootPath, ngmIsEnabled, ngmHostname, ngmMaxClients, ngmUser, ngmPass, ngmKey )

846c846
< (‘DefaultMember’, ‘“._(“Auto generated fog nfs group member”).”’, ‘1’, ‘1’, ‘/images/’, ‘1’, ‘" . STORAGE_HOST . "’, ‘10’, ‘" . STORAGE_FTP_USERNAME . "’, ‘" . STORAGE_FTP_PASSWORD . "’ )",

           ('DefaultMember', '"._("Auto generated fog nfs group member")."', '1', '1', '/images/', '1', '" . STORAGE_HOST . "', '10', '" . STORAGE_FTP_USERNAME . "', '" . STORAGE_FTP_PASSWORD . "','' )",[/CODE]

In these insert SQLs you must put the column names. In the last one, you must add the ‘ngmKey’ column

[warp]

[quote=“Tom Elliott, post: 24338, member: 7271”]I’m still confused. Is Storage_NAS the master node in the Default group?[/quote]
To put it simply, when Storage_NAS is the master node in the Default group, the scheduler say: I don’t appear to be the group manager, I will check back later.

[Tom Elliott]

This is expected. The reason for it to “I don’t appear to the be group manager.” Is the IP of your fog server is not the same as the IP of the NAS. So it is NOT the group manager for that particular node. For everything else, however, it should be perfectly fine.

[Tom Elliott]

r1348 released.

Add’s hooking to the Storage Management Page. Finally, only one more page left.

[Tom Elliott]

r1349 released.

All Management pages are now rid of excessive php tags and table information is all hooking able.

[warp]

[quote=“Tom Elliott, post: 24344, member: 7271”]This is expected. The reason for it to “I don’t appear to the be group manager.” Is the IP of your fog server is not the same as the IP of the NAS. So it is NOT the group manager for that particular node. For everything else, however, it should be perfectly fine.[/quote]
Hello Tom,

Can you then tell me the proper way to set FOG knowing that:

[U][B]FOG server:[/B][/U]
[LIST]
[]CentOS 6.5
[]5 GB disk
[*]FOG 0.33B
[/LIST]
(No backup on the server that has a 5GB disk)

[U][B]The NAS server:[/B][/U]
[LIST]
[]SYNOLOGY
[]8TB disk
[/LIST]
NFS share as many NAS trade is prefixed to the volume name, for it is SYNOLOGY / volume1 / for IOMEGA is “/ nethdd /” which gives for SYNOLOGY “/ volume1/images” and IOMEGA "/ nethdd / images "

I also want to keep the timer function.

[Tom Elliott]

r1350 released.

Major changes with this. (Sort of).

functions.includes.php is no more. I went through and made the necessary changes where needed and was able to remove the need for this file entirely.

Removes the management/includes directory and files as everything is now class based.

Removes the mobile/includes directory and files.

Moves the mobile files to class files.

Sounds like it’s not much, but believe me it should have quite a large impact on overall number of files.

Thank you!

[warp]

[quote=“warp, post: 24403, member: 22860”]Hello Tom,

Can you then tell me the proper way to set FOG knowing that:

[U][B]FOG server:[/B][/U]
[LIST]
[]CentOS 6.5
[]5 GB disk
[*]FOG 0.33B
[/LIST]
(No backup on the server that has a 5GB disk)

[U][B]The NAS server:[/B][/U]
[LIST]
[]SYNOLOGY
[]8TB disk
[/LIST]
NFS share as many NAS trade is prefixed to the volume name, for it is SYNOLOGY “/volume1/” for IOMEGA is “/nethdd/” which gives for SYNOLOGY “/volume1/images” and IOMEGA “/nethdd/images”

I also want to keep the scheduler function.[/quote]
I like to continue to help you to “debug” the 0.33b version, but without an answer to my question, I can go further, because my architecture is that.

[COLOR=#00ccff][B]Can you help me ?[/B][/COLOR]

[warp]

on r1350 there is no page on url “report” : [url]http://srvfog/fog/management/index.php?node=report[/url]

[Tom Elliott]

[quote=“warp, post: 24465, member: 22860”]on r1350 there is no page on url “report” : [url]http://srvfog/fog/management/index.php?node=report[/url][/quote]

This was because there was a function call that didn’t exist anymore. Corrected and you should now see a display of the files and such.

[quote=“warp, post: 24463, member: 22860”]I like to continue to help you to “debug” the 0.33b version, but without an answer to my question, I can go further, because my architecture is that.

[COLOR=#00ccff][B]Can you help me ?[/B][/COLOR] :)[/quote]

The answer to the question, as I understand it, is if you want to manage information on the NAS node, the IP information needs to be different from the FOG Server when in another group. The only issue with this is you won’t be able to Replicate images to this node because it’s no longer a part of that group. For that same reason, putting the node into the group itself won’t work, because the Management of that node is done from a totally different IP. What I’d recommend is to mount the NAS as /images on the local FOG Server. Or under another name if you need (/images) and when you create the node on the Server, you’d still use the FOG Server’s IP address.

So,
Add the NAS to mount on bootup in your fstab file. How you do it is up to you. I’m not going to write a tutorial on how to do this as there’s plenty of howto’s all over the place.

Reboot your FOG Server and verify the NAS is now mounted like a local storage place on your FOG Server.

Add the new folder as a node. The IP address for this node will then be the same as your FOG Server, not the NAS’s IP.

The Management User and Password will also be the same as your Default node.

Make your NAS the Master for that group. (If you have images on the original master, before checking that the NAS volume is master, copy the original to the NAS storage location.)

[Tom Elliott]

r1351 released.

Fixes the report node issue warp reported earlier.

[Tom Elliott]

r1352 released.

Fixes mobile display issues on first load. If you specificed ?node=homes you’d be fine, but if you just went to the link it was trying to display the normal web gui’s home page. This is now corrected. Also fixes the search bar on the hosts node in the mobile page so it’s not a button.

Thank you,

[warp]

[quote=“Tom Elliott, post: 24466, member: 7271”]This was because there was a function call that didn’t exist anymore. Corrected and you should now see a display of the files and such.

The answer to the question, as I understand it, is if you want to manage information on the NAS node, the IP information needs to be different from the FOG Server when in another group. The only issue with this is you won’t be able to Replicate images to this node because it’s no longer a part of that group. For that same reason, putting the node into the group itself won’t work, because the Management of that node is done from a totally different IP. What I’d recommend is to mount the NAS as /images on the local FOG Server. Or under another name if you need (/images) and when you create the node on the Server, you’d still use the FOG Server’s IP address.

So,
Add the NAS to mount on bootup in your fstab file. How you do it is up to you. I’m not going to write a tutorial on how to do this as there’s plenty of howto’s all over the place.

Reboot your FOG Server and verify the NAS is now mounted like a local storage place on your FOG Server.

Add the new folder as a node. The IP address for this node will then be the same as your FOG Server, not the NAS’s IP.

The Management User and Password will also be the same as your Default node.

Make your NAS the Master for that group. (If you have images on the original master, before checking that the NAS volume is master, copy the original to the NAS storage location.)[/quote]
Thank you for the reply,

I actually read all the tutorial for several weeks.

But at the system level, it is impossible to mount a share “NFS” on “/ images” folder which is itself already shared by NFS “exports” file.
[CODE][B]etc/fstab[/B]
172.16.8.43:/volume1/images /images nfs rsize=8192,wsize=8192,timeo=14,intr
[/CODE]

[CODE][B]etc/exports[/B]
/images *(ro,sync,no_wdelay,insecure_locks,no_root_squash,insecure)
/images/dev *(rw,sync,no_wdelay,no_root_squash,insecure)
[/CODE]

[CODE]
[B]sudo service nfs restart[/B]
Arrêt du dÃmon NFS : [ OK ]
Arrêt de NFS mountd : [ OK ]
Arrêt des services NFS : [ OK ]
Shutting down RPC idmapd: [ OK ]
DÃmarrage des services NFSÂ : exportfs: /images/dev does not support NFS export
exportfs: /images does not support NFS export
[ OK ]
DÃmarrage de NFS mountd : [ OK ]
DÃmarrage du dÃmon NFSÂ : [ OK ]
DÃmarrage de RPC idmapd : [ OK ]
[/CODE]