@Wayne-Workman said:
If you want to use 1.2.0, the latest Debian and Ubuntu and CentOS and all late Fedora releases are totally off limits. And I’d recommend an older CentOS version in that case.
Oh. I’m using 1.2.0 on Ubuntu Server 14.04.2. It’s being a pain in the ass. This explains why. What’s a good way to migrate it to Fedora 22 server?
@Wayne-Workman said:
If you’re rebuilding from scratch, I’d recommend moving away from Ubuntu and going with either Debian or CentOS or Fedora Server. However, that choice is up to you… and we will provide help with Ubuntu should you need it.
Which of those 3 would you suggest the most?
I’ve only run into this on one image, but it happens when uploading from multiple hosts. I’ll upload the image, then download it to another and it’ll say “a disk read error occurred”. This happens when I’m downloading to either a host of the same hardware or a virtual box.
Edit: I also find it noteworthy that when downloading to a physical host instead of a virtual one, I get video corruption after the download. As in, there’s oddly colored random characters placed in various places around the screen.
@Wayne-Workman
Okay. I don’t use scheduled tasks. I think it’s wrong because I looked at the log and I was not in the office imaging until 2100 last night. How do I set it? I’m on 1.2.0
@Wayne-Workman
I thought so, but I don’t know how to check the time FOG is set to.
I have my distro syncing with a public NTP server (ubuntu 14.04 x64). While this is a little late to be asking this question, will FOG be affected by the leap second tonight? Oh, well. Guess I’ll find out in the morning.
Can you assign the host an image, try to quick image (not selecting it in quick image)? This should tell whether it’s an issue with all of quick image or just if you select it with quick image.
@Wayne-Workman said:
I don’t sysprep… However, on that same note… my co-worker was messing around with Windows Deployment Services which requires sys-prep.
I remember him getting aggravated about the 3 reboot limit thing… I told him, ok build the image over again… and we will just upload it to FOG for safe keeping. If you need to restart, we can restore it (THEN) in 15 minutes.
And now, we’ve ditched WDS in favor of FOG, and because of the awesome compression and upload / download times in FOG Trunk, we often find ourselves doing a pre-upload of images lol… Uploading what we have before we do something risky that you can’t come back from cleanly.
Once we verify that what we did works, we delete the early image of course. But, we have reverted to early images a few times.
Yeah, I found the speed difference in snapshot vs. upload is well worth building it in a VM, but I suppose if that’s your style then go ahead.
What aer the advantages of building images in VMWare? I know that some people do it but I never understood why… just because you can??
The main advantage, and why I do it, is because you can take a snapshot BEFORE you run sysprep, so that if something is wrong with your unattended file, or you want to update the image and sysprep it again, you can as many times as you want by reverting back to that snapshot. If you don’t, you are limited to sysprepping 3 times per windows installation. It took me a few times to get my unattended.xml right, and being able to update once, push for all is big. My actual method is to build the image, with all the software, in VirtualBox, in audit mode. Then I sysprep /generalize /audit to sysprep it and have it go back into audit mode. Then I upload it to fog from the VM and download it to a physical machine in my office. I boot that one up (into audit mode) and install all the drivers I need, then sysprep again with the answer file “sysprep /oobe /generalize /unattend:unattended.xml” and upload, then multicast to the rest of the physical machines.
the client reports all available mac addresses when checking for tasks. this means that, typically, the registered mac is sent with the pending tasks check whether it’s the one being used to connect to the network or not.
That’s what I was wondering. Thank you.
In SVN
That means not in 1.2.0, right?
I’m imaging a bunch of netbooks. Once I disconnect them from the ethernet, will they still communicate with FOG okay, or will they not because they’ll be referencing the wireless MAC only? Do I need to add the wireless as an additional MAC? Is there a way to do so automatically so I don’t have to type in the MAC addresses for 100+ netbooks?
Yeah. I don’t know how I missed that.
So I need to do:
sudo su -
cd /var/lib/mysql
mv ibdata1 ibdata1.bak
mv ib_logfile0 ib_logfile0.bak
mv ib_logfile1 ib_logfile1.bak
cp -a ibdata1.bak ibdata1
cp -a ib_logfile0.bak ib_logfile0
cp -a ib_logfile1.bak ib_logfile1
service mysql restart
and
echo manual | sudo tee /etc/init/mysql.override
Edit the /etc/rc.local file before the exit 0; line add:
sleep 30 && start mysql
right?
Okay, so he said:
One of the things I hate about Ubuntu, lately, is exactly this.
Ever since moving to Upstart to control services, the services seem to no longer be controlled.
It will start services before the dependent services (Networking, Local system, etc…) is running.
This breaks other things as well, such as the linux side FOG Services (FOGMulticastManager,FOGScheduler, etc…)
What’s happening, from what I can see, is it starts mysql before networking is ready. It works for a small amount of time, but then when networking is ACTUALLY ready, the service fails. Upstart is supposed to handle restarting a crashed service, but it’s almost like it was completely forgotten about.
But I don’t understand, am I supposed to find or know a way to delay or rearrange the order the servies start up?
Sorry. Thought I put that in there. Ubuntu server 14.04.2 x32 with fog 1.2.0
When I first boot my server, I have to start MySQL manually and also stop, then start FOGMulticastManager manually. If I run service --status-all before I do anything with FMM, it has a question mark. If I try to start it, it shows Fail. If I stop it, then start it, it works.
Edit: Running FOG 1.2.0 with Ubuntu Server 14.04.2 x32