@tutu10 i think you posted this to the wrong forum…

Did some more testing and will merge this into dev-branch now before the new release. Have tested on CentOS (existing installation and fresh) as well as Debian (fresh installation). Should be good to go.

@tejen While @ckasdf is totally right about saying that it sounds like you might be using FOG as a backup tool and it’s not a great idea, I can still tell you that you can create scheduled jobs. Just do as if you’d manually create a capture task but on the last step instead of do an “instant task” click schedule as cron task.

@klaus-jauk In FOG groups are not like you might be used from other software. It’s more like a tool to push out settings to a group of hosts but it’s mostly not a persistent settings thing where you store information for the whole group. For AD settings, if you set those through the group you’ll see that the values will be shown in the group AD settings as long as you don’t change a value for any of the single hosts. When you do this the value from the group AD settings will disappear.

This is the concept of FOG groups. Search the forums for “persistent group” and you’ll find a lot of information on this.

I see that it’s a bit unfortunate in the case you describe where you register a new host and join that to the group. But on the other hand you’ll still have all the values set for all the other group members. You just don’t see it in the group AD settings view anymore.

@Sebastian-Roth Okay, I will give it a try. Thanks for the response.

@vervoordeldonk What we are finding is the mainstream hardware vendors (dell, hp, lenovo, intel nuc so far) do an excellent job of making each system unique so coming up with an alternate method of identification isn’t a problem.

The issue is like the system below

System Manufacturer xxxxr System Product TMB-3010 System Version R0M0 System Serial Number To be filled by O.E.M. System Type Type: Notebook BIOS Vendor American Megatrends Inc. BIOS Version 1.00/636G00WPC76508 BIOS Date 05/06/2016 Motherboard Manufacturer xxxxr Motherboard Product Name WEB-889A/WEB-889AA Motherboard Version R0M0 Motherboard Serial Number To be filled by O.E.M. Motherboard Asset Tag To be filled by O.E.M. CPU Manufacturer Intel® Corporation CPU Version Intel® Core™ i5-3210M CPU @ 2.50GHz CPU Normal Speed Current Speed: 2500 MHz CPU Max Speed Max Speed: 3800 MHz Memory 3.74 GiB Hard Disk Model 2.5" SATA SSD 3ME3 Hard Disk Firmware S15A19 Hard Disk Serial Number xxxxxxxxxx000025 Chassis Manufacturer xxxr Chassis Version Chassis Serial To Be Filled By O.E.M. Chassis Asset

I put the lower case “x” in just to mask a few values, but as you can see there is nothing unique on this system from the next system from the same manufacturer other than the disk serial number. I have a few examples of other systems that are very similar. One solution is to go ahead and use the mac address as part of the mix for system identification. In that case we would be no worse off than we are today by only using just the mac address.

We also want to bring in the asset tag value so at the very least if FOG can’t identify the system the IT tech will need to populate that field to create a unique ID for that system. (sorry I’m a bit thinking out loud here)

I’ve updated the rout.class.php in 1.5.X branch (dev-branch)

I will ask @MarkG if you would like to install this to see if it address what you need directly?

I know you don’t want to move to the alpha build, but dev-branch is where the bug fixes found since 1.5.5 have been addressed.

If anything, you can simply copy this dev-branches router class with:

wget -O /var/www/fog/lib/router/route.class.php https://raw.githubusercontent.com/FOGProject/fogproject/489ba2952f5750e1483083eedade17dde6bb3cd8/packages/web/lib/router/route.class.php

This should allow you to clear out items such as snapins with the code you were using:

{ "snapins": [], "printers": [1,2,3,4] }

@Sebastian-Roth

No, unfortunately not. I was so busy these last few days that couldn’t make some time to experiment.
As I said, I just took over the management of our university networks and setting everything right will take some time.
But I have this issue on my to-do list because O am directly interested so I didn’t forget. I’ll get back here when I have news.

Update

Still unable to get the MACs from Host(), but I think I’m getting closer. I just dont know what I’m supposed to do with the output of a ->get() call.

public function onEvent($event, $data) { $hostData["name"] = $data['HostName']; $iData = var_export($hostData,true); $hostObj = new Host($hostData); $hData = var_export($hostObj,true); $primaryMac = $hostObj->get('primac')->get('mac'); $macList = array(); foreach ((array)$hostObj->get('additionalMACs') as $ind => &$MAC) { $macList[] = $MAC; } $oData1 = var_export($primaryMac,true); $oData2 = var_export($macList,true); file_put_contents("/var/www/fog/ds_fw.log",$iData . "\n" . $hData . "\n" . $oData1 . "\n" . $oData2); } }

$primaryMac results in NULL, and $macList results in an empty array. I picked/adjusted the $macList loop from one of the pages code, so that ‘should’ theoretically work, as long as the get()'s actually return data.

@CWDS Not sure why it would go through the interfaces twice.

Anyhow, if you want to change the behavior for iPXE you want to look into building your own iPXE binaries and play with the ipxe script: https://github.com/FOGProject/fogproject/blob/master/src/ipxe/src-efi/ipxescript

For iPXE building you wanna read here: https://wiki.fogproject.org/wiki/index.php?title=IPXE#Compile

It’s been a while since I wrote that code and I’ve been working hard on trying to get 1.6 to a more functional state (which I believe it’s mostly there now.)

Of note, while you can tie into the common ajax code sitting within for setupTimeoutElement and submithandlerfunc, the common thing to be noted is that you can have your JS code do whatever you need. Those methods are available, but there’s no 100% dedication to using it. Use it where you can, but it’s not a requirement.

The js code get’s loaded automagically in 1.5 via node and sub in the get request.

So let’s say you wanted js code to alter the list page for some reason. Your plugin node code would be say exampleplugin. You would simply need a file named: fog.exampleplugin.list.js (file names are always: fog.<node>.<sub>.js)

Hopefully this makes sense.

So in keeping with that file scheme, common files I can think of for a plugin named exampleplugin would be:

fog.exampleplugin.js //base js for all exampleplugin node items. fog.exampleplugin.list.js //exampleplugin's list page. fog.exampleplugin.add.js //exampleplugin's create/add item page. fog.exampleplugin.edit.js //exampleplugin's edit item page.

While I’m sure this doesn’t answer any of your questions, it should help with getting something functional.

You can use other plugin’s to help in the understanding of the js files too. A common one I like to use that interacts with both groups and hosts is the location plugin. This should give a pretty good understanding of what to do, and how to do it. If you still have questions please ask and try to be detailed in exactly what you’re looking for.

@Tom-Elliott ok thank you. We are starting with doing only some basic things, so I will just continue to get it up and running on 1.5.x.

Right now, I am using the doxygen generated focdoc/ folder and mostly some examples in other plugins. Sometimes a bit tricky, but IMHO no need to generate the files for 1.5

@klaus-jauk I understand for $osid and driver packs. In the case of drivers, using FOG’s built in $osid is not a problem, because for Win10 for example a win10 drive works across all win10 platforms. The same for Win8 and so on.

Here is an idea: What I see is at issue is the KMS key where you think you need custom $osid values. There is a key management plugin for FOG. Where you load the KMS/MAK keys into a key table and then associate that key with a host. I don’t know if that will solve your problem with kms keys or not, but the resultant link key should show up in hostproductkey so you don’t need to do anything special in the post install scripts.

The issue I see with creating custom $osid values is that FOG uses that $osid values internally when creating the disk partitions and how to deal with boot block. If you create your own $osid values then the internal programs that look for $osid (9) for example would not work because it would see your custom $osid value. Now of FOG internally would mask off (only look at) the lower 8 bits of $osid for FOG’s internal use then you could do what you want by creating custom IDs with the upper 8 bits of that $osid value.

@Fernando-Gietz said in Use an external database:

My FOG version is 1.5.5 for working branch, the last version.

So you don’t have the fixes I pushed to dev-branch. We might consider merging your code changes form working to dev-branch soon. This way you can make use of all those changes.

@Tom-Elliott What do you think?

@Fernando-Gietz said in Performance Monitoring tools:

We use Nagios to monitoring the server status but not use any plugin to see the activity of MySQL or Apache. Do you use any plugin to monitoring these services?

We use plugins for both mysql and apache monitoring. But I feel I must make a distinction between performance monitoring and service level monitoring. We do service level monitoring, in that we only check if the service is functional (what ever that means like mysql connections, or basic http queries) as well as critical service parameters. We dont monitor things like inbound network traffic for sql server, because its really not important to know how well the sql server is functioning (I know you can make an argument either way). As you know nagios is just a go/no-go monitoring system. If you couple that with mrtg or such then you can get trending over time. That is why we really like Centreon because it has nagios coupled with a few other opensource softwares in one package. You can get that trending over time of the actual recorded value as well as the go/no-go status.

Performance monitoring is collecting a large amount of metrics to understand how the system as a whole is functioning, not just the critical bits. You could use performance monitoring environment for forensic investigations over time to understand how metrics have changed. You may be not so interested in the individual services here.

So when looking for a monitoring tool, you need to decide what is important about that device/service. Why does it exist? What function does it perform? What would happen if that service disappeared? That is how you should decide what to monitor regardless of the tool.

For example: Lets say you wanted to monitor your FOG server for both availability and performance metrics. What do we know about your FOG service? Well it runs on linux, it uses a tftp server, apache server, a ftp server, and nfs.

So for linux what is important?

Is it reachable on the network? How utilized is the CPU? How much free memory is there? How much free disk space is there on the /images disk? Are the tftp, http, ftp services running in memory? (notice I didn’t mention if they were working only if they are running in memory)

For the tftp server what is important

Is it listening on port 69? Does it respond to a file size request?

For the http server what is important?

Is it listening on port 80 Does it respond to http requests? If your site uses php does it respond properly to a php page call? What is the percentage of CPU does the apache service use?

And so on for monitoring. The point is to understand what you want to inspect before you pick the tool. Then find the tool that fits your budget and time you are willing to invest in setting it up.

At my brothers company they use a product called WhatsUp Gold and they loved it before they moved to PRTG which they now love more. But they had the budget but not much time to build it themselves. In my case I’ve been running nagios monitoring servers on and off since 2000. I don’t have a problem spending time to build new installs.

@ragnurenson so, I’ve just tested after applying the commited changes and everything works fine! Thank you for pointing out the github commit and thanks @Fernando-Gietz for the awesome work!

If I can make a suggestion, I’d say it would be nice to have that search filter implemented in “List All Hosts” screen on Active Tasks!

I made the changes and did a test run, dev-branch completely passed.

Unless, of course, you’re trying to make it core usable? But the only time this would be viable would be if LDAP plugin is in use. Of course other plugins may want to make adjustments here as well. I need to think on this a little bit.