Posts made by EduardoTSeoane

@Fernando-Gietz @Sebastian-Roth

I’m trying to enable multicast on my installation. And I have a problem.

Trying to change nic on node definition or in fog settings dont change on the multicast command.

FOG try to send the stream always from the eth0.

All my nodes are multihomed with at least 3 VLANs, and eth0 is a management interface on the management network.

I try to locate where the task get the nic to use but i can’t find it. Well i find some places but forcing manually by nic name and by nic ip dont works.

Testing manually udp-sender and udp-location is multicast capable.

The installation are Normal and Node v1.5.6

Thanks in advance.

i’m doing that with only an ip, but loosing multicast., no special configurations needed if you have access to all services on the server. Maybe nfs ports are blocked on the router/firewall?

@Sebastian-Roth yes, can be closed thanks.

@Wayne-Workman , @Fernando-Gietz , @Sebastian-Roth

I’m thinking about your post and I’ve done a revision of some computers’ history.

And I can confirm that i have snapin sequences, that i know that they are executed on launching time order, not in alpha-numeric.

BUT this has been caused by a procedure change.

After reviewing the procedure and some tests I have noticed that:

• a snapin sequence launched to computers is executed in alpha-numeric order

• a snapin sequence launched to groups is executed in launching time order

Thank’s for make me think a few.

@Sebastian-Roth
Upgraded to 1.5.6 as you suggested.

It seems to be a bit faster, but i need to do the corrections on the server with the same values.

Now i’m experiencing another issue posted as https://forums.fogproject.org/topic/13328/fog-snapins-not-ordered-on-launching-time

@Fernando-Gietz , @Sebastian-Roth .

Recently updated to FOG 1.5.6.

From the update from 1.4.2 to 1.5.5, and now updating to 1.5.6 we have the following trouble.

The clients are getting the launched snapins ordered by name, not by launching time.

I get this error with and without location module.

the php dont request the needed memory, it request for the next block. So:
When i has 256M it request another ±20M.
Then I increase to 512M. Then the system claims for another 20M.
Then I increase to 1024M and then the system claims for another 20M

I repeat the actions to accure memory size to 1536M.

If it’s necesary to increase the memory a lot then maybe it’s necesary increase the TimeOut and ProxyTimeout httpd.conf paramenters.

Promised Example:

## Elevate to admin the current Script Execution
If (-NOT ([Security.Principal.WindowsPrincipal][Security.Principal.WindowsIdentity]::GetCurrent()).IsInRole([Security.Principal.WindowsBuiltInRole] "Administrator"))
{  
$arguments = "& '" + $myinvocation.mycommand.definition + "'"
Start-Process powershell -Verb runAs -ArgumentList $arguments
Break
}
##Create credentials
$password = ConvertTo-SecureString 'MySecretPassword' -AsPlainText -Force
$user='useradmin'
$credential = New-Object System.Management.Automation.PSCredential ($user, $password)
##Execute wusa.exe with parameters and useradmin credentials.
Start-Process -FilePath <path wusa.exe> -ArgumentList <arguments for wusa.exe> -Credential $credential -LoadUserProfile -PassThru

This maybe works

With New-PSDrive you can mount samba and nfs shares if you have the nfs-infrastructure installed.

@fry_p I use 1536M to list 15.225 pc’s.

Little question,
Have you restart php-fpm and httpd services?
I forget to do it when i did the change and…

Understand, isn’t it?

this is in /etc/php-fpm/www.conf
the parameter is php_admin_value[memory_limit] =

@Sebastian-Roth we currently have 20 or 30 at a day, and this year we go to renovate a big percentage of the park.

Sorry approve pending hosts.

As I say on the title.

It’s good to can do groups to approbate pc’s in block.

I had that problem on some notebook models.
In my case was the boot file.

Changing the undionly.kkpxe that im using (on the standard boot files) by the realtek.kpxe it works.

Try the different boot files. Maybe some works.

Now my dnsmask and my isc dhcp deliver different file based on mac vendor bytes.

@Sebastian-Roth @Fernando-Gietz
I’m glad to help in everything i can, only have to request.

Tomorrow I do a test, to start, on my dev environment with the same database.

By the moment I put the following feature request:

https://forums.fogproject.org/topic/13295/move-the-node-list-report-from-about-page-to-another-page-linked-on-about-page

It’s good to have this page, but in some cases, with a largest list of nodes, an if you have some nodes stopped by maintennance, can be to slow to load.

I think that is better not to have to wait for that list to go to fog settings page.

It’s not the same wait for a page that you don’t want to visit, than wait for a page that you want to visit.

@Sebastian-Roth @Fernando-Gietz

Solved!!!

Solution:
In /etc/httpd/conf.d/fog.conf (outside and before of VirtualHost tag):
Add this parameters:

• TimeOut 120
• ProxyTimeout 120

In /etc/php-fpm.d/www.conf

• Increase php_admin_value[memory_limit] = 1350M

But it’s hard to load the info.
Thank’s to all for help

To be exactly the NDR is HTTP ERROR 500

@Sebastian-Roth

rpm -qa | grep php
php-common-7.2.18-1.el7.remi.x86_64
php-7.2.18-1.el7.remi.x86_64
php-ldap-7.2.18-1.el7.remi.x86_64
php-fpm-7.2.18-1.el7.remi.x86_64
php-json-7.2.18-1.el7.remi.x86_64
php-cli-7.2.18-1.el7.remi.x86_64
php-mysqlnd-7.2.18-1.el7.remi.x86_64
php-mbstring-7.2.18-1.el7.remi.x86_64
php-pecl-zip-1.15.4-1.el7.remi.7.2.x86_64
php-process-7.2.18-1.el7.remi.x86_64
php-pdo-7.2.18-1.el7.remi.x86_64
php-bcmath-7.2.18-1.el7.remi.x86_64
php-gd-7.2.18-1.el7.remi.x86_64

@Fernando-Gietz
yes at the moment i found apache general TimeOut and fast-cgi ProxyTimeout.

That solve the storage list (node=storage) and the list of nodes in about page (node=about).

Maybe i must increase a bit more to test, but the error page is diferent.