Posts made by Jim Graczyk

Upgraded to the latest working version.

Still cannot delete an image from a storage group. I have the location plugin installed. One node per location, different storage group at each location.

Jim

Upgraded to the latest working branch version.

The Image Log (under Reports) now works and appears to display correct data.

The Snapin Log (under Reports) still shows “No results found”.

Also found that the User Tracking Report has problems that were also seen in 1.5.0.RC2. I think the host name isn’t making it to the report. When you run the report with just a host selected, the list shows usernames that have logged in but the first column, host name is blank. When you run it with just a username selected, get only one row with the host blank and the user equal to the selected user. When you select both a hostname and a user name, you get only 1 row with no hostname and the username you selected.

I would expect this report to list host, user and a timestamp for all records reports choices.

Jim

@Tom-Elliott said in Storage Group in Dashboard Showing 1 Slot in QUEUE But No Tasks Exist:

mysql fixing

I ran the maintenance script found at the bottom of this link:
https://wiki.fogproject.org/wiki/index.php?title=Troubleshoot_MySQL

Several lines deleted rows. The problem is resolved.

Thanks,

Jim

Tom,

I regret that I accidentally upgraded to 1.5.0 RC2 - it was stupidity on my part - I only found the wiki to the ‘latest’ FOG release. What does one ‘checkout’ to get the working releases? How does one know when a new working release is made?

Thanks,

Jim

Joseska,

I suggest a snapin would be a better place to add the computer to a group in LDAP. Assuming the computer is a Windows computer and you want to add that computer to the Samba emulation of Active Directory and assuming Samba emulates AD such that Windows PCs can find the Samba system(s) like they would a real Microsoft Domain Controller, a CMD command would do the trick. I’ve used commands like this below to put domain groups into local computer groups. This command, run as a snapin on a PC in a domain, will add a domain group to a local computer group - giving users in the domain group rights on the PC:

net localgroup administrators DomainName\DomainGroupName

I also believe you can add an AD computer object to a domain group with the same command. It’s abnormal to place a computer object in a domain, but I’ve used this approach to allow services on a specific PC to access network resources without having to place an AD account on a service. The down side of doing this is that all services on that computer have the same access rights, not just the service you intend to have those rights.

I believe this command will add a computer to an AD group:

net group DomainGroupName computername$ /add /domain

I don’t know if your LDAP/SAMBA/AD supports Group Policy Objects, but actions that need to be taken on machines just after they join a domain (so they work correctly for future users) are typically done in Startup Scripts inside a GPO applied to the OU to which the Machines are joined.

Hope this helps.

Jim

There are several screens worth of results to the query (397 rows in set).

Getting a log out is not simple, but if you want a list, I’ll post it.

Several records have stReturnDetails == Pending… and Caneled due to new tasking
stState values include 1, 4, and 5. All records with Pending… and Canceled stReturnDetails have stStatus of 5, but not all redocrd with stStates of value 5 have Pending or Cancelled… stReturnDetail values.

Many records have stCompletDate == 0000-00-00 00:00:00
Very few of the records returned have stCheckInDates after the upgrade to 1.5.0 RC2, so I’m guessing the problem is old, but the reports still returned something in 1.3.4 and 1.4.2.

I have no need for this task history at this time (I will later, but not now). If I need to delete records please provide syntax.

thanks,

Jim

Server

• FOG Version: 1.5.0 RC2
• OS: CEntOS 7

Description: FOG Dashboard Show 1 Item Queued and one Imaging Slot Taken, But No Task Exist.

This may or may not be a 1.5.0 RC2 issue or it could have existed in previous versions. See attached screenshots.

Is there referential integrity regarding active tasks, where they reside and the dashboard display?

Thanks,

Jim

Server

• FOG Version: 1.5.0 RC2
• OS: CEntOS 7

Client

• Service Version: 0.11.12
• OS: Win10 & Win7

Description

Just upgraded from 1.4.2 to 1.5.0 RC2. FOG and 2 Storage Nodes, all on CentOS 7, location plug in installed. Each server in its own location and site. Sites separated by firewalls and static VPN WAN paths. Each FOG server is in its own storage storage group.

I had a bad replica of an image at one site. I attempted to delete the image from that site/storage node with the new UI by selecting the squarish CheckBox and pressing the RED Remove button. The system paused and then posted a dialog box that the image had been updated, but the display remained the same. I then went to a different ‘tab’ on the same image UI (like Members, for example) then back to the Storage Group Tab, but all 3 storage groups remained in the list. While the UI is a bit unintuitive on this screen, it’s clear the squarish CheckBoxes apply to the Remove button given that the only one of the circular RadioButtons can be selected at a time, and the Remove button have ‘Remove Selected Groups’ next to it.

To be sure the problem really exists, I then went to the Members ‘tab’ and removed any host that was on that site, that might be blocking the deletion of the Storage Group from the Image definition. This had no effect. I could not delete an image from a storage group using the UI.

I have yet to try to add a new image and see what happens.

If you need screenshots, just let me know.

Thanks,

Jim

Server

• FOG Version: 1.5.0 RC2
• OS: CEntOS 7

Client

• Service Version: 0.11.12
• OS: Win10Pro & Win7Pro

Description: Snapin and Imaging Reports Failing, Problem with Snapin History in Host Area

Just (accidentally) upgraded to 1.5.0 RC2 on one FOG Server and 2 Storage nodes from v1.4.2. All CEntOS 7. Location Plugin was already installed in 1.4.2 pre-upgrade (maybe was installed initially on v1.3.4). Each Storage node is at a separate Location, in a separate storage group.

Snapin Report is blank / empty while Snapin History on Host lists snapin installs that I know happened. I was anxious to see if the Snapin report got some date and time filtering, but I see only No results found. There may be a data integrity issue. The Snapin History listing in the host area lists blanks in the Snapin column for old and new snapin tasks that have Queued status. This is odd since the v.1.4.2 system had no queued tasks when we took it down to upgrade it.

A side comment - I was hoping the Snap Log file would no longer spit out the complete history of all snapins ever deployed by the server when one initially requests the report (as v1.4.2 did). This change from previous versions of FOG made the report unusable. The list quickly become GIGANTIC, which takes a long time to arrive over slow connections and time-consuming to scroll thru.

I’m receiving an HTTP error 500 on when I select Image Log. The image history in the Host area appears to be correct.

If you need screenshots, I’ll post them.

Thanks,

Jim

Andreas,

/tftpboot is the correct folder for fogexe.ini. If you can run tftp from the PC on which you’re trying to run the snapin, that’s a good sign. I suspect it’s a name resolution issue. FogExe uses the name FOGSERVER as the name for your MAIN fog server. You have to make that resolve to your actual server’s IP. The PC will lookup the name FOGSERVER using the short name, so FOGSERVER has to be defined in the PC’s namespace, not just the fog server’s namespace.
[LIST=1]
[*]Verify you have a CNAME record (alias) for the name fogserver in your DNS for the DNS domain to which your PC belongs. The alias needs to point to your main fog server. If that server’s name is myfogserver.mydomain.xyz and the PC’s name is pc1.anotherdomain.abc, your DNS for anotherdomain.abc needs to be able to resolve the name fogserver.anotherdomain.abc to the ip address of myfogserver.mydomain.xyz. The best way to do this is to have an alias in the abc space for fogserver.anotherdomain.abc = myfogserver.mydomain.xyz and have forwarders in the abc root to the xyz name servers . Another way is to just add an A record for fogserver and put in the IP address of the myfogserver.mydomain.xyz. Test that DNS works by entering the following on the PC trying to run the snapin:

ping fogserver

By fogserver, I mean use the word fogserver as the name. Do not substitute your server name. FogExe uses this as an alias -it’s hard coded. It’s your responsibility to make this name resolve to YOUR server’s IP. If this works, then you can test further.
[]On the PC, go to your FOG folder, in the tmp folder you should have the following files:
FogExe.exe
tftp.exe
wbdxxxxx.dll
wwprc44i.dll
wwwnt34i.dll
and maybe FogExe.log
if you don’t have these files present, you don’t have FogExe on your PC. Run DeployFogExe.exe in that folder or follow deployment for FogExe in the guide.
[]Next we need to test the tftp client FogExe uses. On the PC that you’re trying to run the snapin, go to the command prompt and navigate to a folder you have write access to (like %temp%) and into this folder copy the tftp.exe from your FOG/tmp folder (above). Run this command:

tftp -v -s fogserver fogexe.ini

This is the command FogExe runs to retrieve fogexe.ini. Be sure the file on the fog server is fogexe.ini - all lower case. I’m not sure if the tftp client will care, but I know LINUX does, so use lowercase to be safe.
[/LIST]
These test should show where your problem is. If not, please post FQDNs for your PC and main fog server, your fogexe.ini file and the results of the steps listed here and we’ll get to the bottom of the problem.

Jim Graczyk

Here’s the file I couldn’t upload before. It’s for creating a Snapin to seed FOGExe.exe into the client PC so it can be used by subsequent Snapins.

Jim Graczyk

[url=“/_imported_xf_attachments/0/291_DeliverFogExe.zip?:”]DeliverFogExe.zip[/url]

Scaling FOG to work in a multisite environment has been a topic in this forum for a while. FOG is great, but has some design issued that more or less require separate instances of FOG at each site. Software distribution has more serious limitations, even when used at a single site.

Please find attached a detailed guide to setting up a system of FOG servers that allows for a single GUI/DB node for central administration with site-based distributed storage nodes. Also, the guide shows how to use FOG for central software distribution by leveraging MS DFS to provide redundant site copies of installation files. Executables are provided to facilitate snapin installation from local DFS cells using only 20Kbyte of WAN traffic.

The process has been tested and demonstrated to work with unicast and multicast image deployment. I’m placing this doc and exe’s out for general use to more fully test it’s effectiveness.

Jim Graczyk
Netway Solutions, Inc.

P.S. - I was unable to post the file DeployFOGExe.exe referred to in the guide. The file is 1.3mb and this UI won’t allow me to upload it - says it’s too large. If an admin would assist, I’ll post it later.

[url=“/_imported_xf_attachments/0/288_Scaling FOG for the Enterprise.pdf?:”]Scaling FOG for the Enterprise.pdf[/url][url=“/_imported_xf_attachments/0/289_FOGExe.zip?:”]FOGExe.zip[/url][url=“/_imported_xf_attachments/0/290_FogExeStub.zip?:”]FogExeStub.zip[/url]

I’m new to FOG, but I think your problem is the FTP/TFTP password. This isn’t the password you use to log onto the FOG web UI. FTP and TFTP pass usernames and passwords as clear text. This can be pulled out of a sniffed packet or thru several other means. To prevent this, the password included in the config.php (accessible within the web UI, as well, under Other Information/FOG Settings) is actually the public key string produced from running the password against an encryption algorithm that uses a private key that’s programmed into the FOG system. Should you’re interested, you can change that key to guaranty privacy, but that requires recompiling the FOG system code.

So to generate the text string to enter into the password field, you have to extract a Windows program and at least one DLL from the FOG tarball. I found it to be incredibly odd that they would included a Windows executable in the LINUX file, and it’s not available in the same place from which the FOG Windows client can be downloaded, but whatever…

Here are the posted instructions for using the tool - called FOGCrypt.exe - which I didn’t exactly follow:

[url]http://www.fogproject.org/wiki/index.php/FOGCrypt[/url]

Instead, I downloaded FOG from the FOG homepage and used 7zip to open the file, then found FOGCrypt and extracted it and the included DLL(s) to some temp dir. I didn’t bother to alter the source code with my own private key - I just used the one they include - for now anyway.

Run FOGCrypt.exe against your password and store the program’s output in the password field in config.php.

Jim Graczyk

Thanks for the clarification. I guess I’ll just have to get over my distrust that MS would actually correctly deal with all hardware platforms. This has never been the case and I’ve been doing this for 20 years, so it’s quite a leap for me. I like your solution since it leaves the drivers in the packs, making the image smaller. For mine process, I have to expand them.

Jim

I wish I’d found this guide about a week ago. Thanks to Andyroo54 for the hard work to get it together and the willingness to share it.

I’ve done essentially the same thing, but without the DP_Install_Tool. I do have a fundamental question on the efficacy of the process. I apologize in advance if I’ve misunderstood any of what’s been documented here. Just point out my error. So here’s what I think I understand:
[LIST=1]
[]This whole process is an attempt to make an image that will work on a wide array of hardware platforms?
[]OS is installed
[]drivers are copied to a drivers dir on the machine prior to imaging (driver packs).
[]system is sysprep’d
[]System is uploaded
[]image is sent to some other PC, boots, sysprep runs, reboots, etc
[*]of the first logon on the new machine, DP_Install_Tool runs, with your defined alterations and drivers are installed, unused stuff is deleted.
[/LIST]
If all of the above is true, how does the OS boot the first time if the image knows nothing about critical chipset or mass storage drivers that are normally required to boot the OS? DP_Install_Tool runs after the first boot. It seems like a chicken and egg issue. How can the OS boot in the first place.
I’m sure I’ve got something wrong. If I don’t, the result would be the same as sysprep’ing XP without [OEMMassStorage] and all the effort that required to extract PnP IDs. In this scenario, a freshly imaged PC would deploy correctly, all drivers would load for new NICs, sound cards, graphics cards, etc., if and only if the chipset (almost all were) and the disk subsystem were supported by whatever was on the original MS OS CD (often a problem).
The process I’ve just started to use I believe solves this problem by “injecting” drivers into the image using DISM from a Winbuilder or WinPE boot CD. This is done between steps 4 and 5 above - post sysprep, pre-upload. The step uses the same driver packs, but extracted to a second ISO or run from a network share. Getting DISM to work correctly in a Winbuilder CD was the tough part. This step seems to require the OS drive not be in use.
If the guide give us a process that works for all devices, including mass storage, I’d rather us it than the one I’m using since it adds a different type of support cd and make the process a little more difficult. I’m just getting started on this, so the right direction really matters
Please advise.
Jim Graczyk

Bill -

I’m new to FOG, a complete newbie on LINUX, but fairly well versed with the MS world. I have just come up with a way to enhance FOG snapin deployment using DFS. I had assumed the rsync stuff I’ve seen would get me replicas of the images on different storage nodes, but I want redundancy and multiple sources at each site for snapins. Further, I find the whole single-file installation limitation of FOG untenable. I can’t have software installation content stuffed into a single file, as most SW in the MS world consists of several files.

So I wrote a program called FOGExe.exe. It leverages the TFTP site where a FOGExe.ini file is stored that has the sharename (DFS or UNC), username and password for accessing the Windows server. A parameter in the snapin parameter field points to the dir within the share and the program looks there for a specific .cmd file - FOGExe.cmd. This cmd has all the commands you’d want to perform the snapin installation. I typically put two versions of an installation in such directories - one to install and the other to repair.

This program is brand new and has been only lightly tested. As I’m new to FOG, I can’t say I won’t enhance this more over the next little while. I’m using FOG now as a replacement for a homegrown system that used BartPE for XP and WInBuilder for Win7 to create ISO that were PXEbooted and uncasted images with Ghost32. For software deployment we had a system that could detect if someone was logged in and would allow for interactive and silent installation. I may want to retain this feature and it would be easy to add it to FOGExe.

If this is of any interest to you, I’ll post it sooner rather than later. I had planned to post it but as I said, it’s new. I’d appreciate any info related to your problem with syncing Linux storage repositories over WAN and how one mounts DFS on a Linux box. Also, why use the Linux bos for TFTP at all? If you can copy the contents of the linux tftpboot dir to windows? I’ve used pxelinux.0 and pxeclient.cfg on windows for some time.

I’m not sure how these forums work, so I’ll check back here for a response.