Posts made by gwhitfield

Understood. I’m just saying I had this happen when I created an imaging task for a computer without having an entry in the “Host Image” field. It was an older version of FOG at the time, I don’t know which one you’re using or if this is even possible with the newer versions.

To OP I found that setting PIGZ to 0 actually slowed my transfer rate down significantly. I currently have it set at 3 and we’re happy with 5GB/min downloads. I can imagine setting it at 6 may also be slower than 3 though I haven’t tested that. I also know that NIC driver configs have everything to do with the speed though I know you said same OS so maybe there’s nothing there, maybe.

Happens to me when I forget to specify the image to use. Not that it happens alot, but…

I’d like to piggyback on this. We have been using FOG for a number of years. Prior to that we had Zen and Ghost. We work with approximately 1000 PC’s of constantly changing makes and models spread throughout 15 or so schools and a number of additional offices connected (or not) to those schools. Our very small team of five doesn’t have time for steep learning curves or to tweak and play and optimize and we are all Linux sub-noobs, yet FOG has always managed to do what we absolutely needed it to do. There is much more to FOG than we understand how to use but even so with the help of a great community the capabilities we’ve gained make FOG something we can’t do without. I think that speaks highly of the quality of the product and the amazing skills and commitment of the individuals developing it. THANK YOU ALL!!!

I know it’s an old thread but I’ve looked at others and didn’t get any closer to an accurate description of the problem we’re struggling with and now may have solved for ourselves.

We have server (A) FOG 1.2 on Ubuntu 14.0.4.1 hosted by vCenter Server 5.5.0. 2CPU/4GB RAM/e1000 NIC/SCSI drive. With PIG_Z_COMP set at 3, upload speeds were no better that 500MB/min while download speeds were in the 3GB/min range. After much googles I find some references to VMWare driver issues solved by command line “(sudo) ethtool -K eth0 tso off gso off” which worked. Upload went to 4.5GB/min on 1Gb connection, download went to 5.1GB/min without a restart of FOG server. There’s also an rx-checksumming edit suggested but we didn’t implement it to get to the speeds we’re at now. Maybe we can live without it.

Funny, we have another server (B) FOG 1.2/Ubuntu 14.0.4.1 VM 2CPU/4GB RAM/e1000 NIC/SCSI drive on ESXi 5.1 that required no changes to achieve 4.3GB/min uploads on 1Gb connection.

Now to get handle on how to make this change stick after a restart, something in /etc/network/interfaces but nothing worked yet, maybe a syntax thing idk. Also trying to get mysql service starting at the right time to avoid the Database Schema issue on restart. I know it’s in here somewhere, just haven’t found it yet. When I do I’ll try to update this.

Thanks, Tom, I hadn’t come across the MySQL thing but I’ll make sure we’re doing the start-stop thing right from here on out.

Pittsburgh, Pennsylvania
USA

Guess I could be more familiar with the product. We’re no experts but we can create a FOG server that works so we’ve just gone with what was simple and worked up to now, avoiding cross-subnet and WAN traffic at all costs. Would storage nodes really help in this situation since we don’t have a need to replicate anything to the master node or to any other node? Every location is a school with widely varying image requirements. I’d say between ALL our locations we probably have 40-50 different images, too much for any one repository.

I’d like to clone a FOG server (without images in it) and deploy it just like I do a Windows image. I have probably 10-15 FOG servers, almost all Ubuntu 10.04 and FOG 0.29 throughout various remote locations and building new ones from scratch every time seems silly considering this is an awesome cloning solution. I have a bunch of identical retired Dell 755 desktops with big hard drives so hardware differences wouldn’t be an issue. I know I’d have to make some minor config changes with IP’s, etc but I can’t find anywhere that this is discussed. How to do this? Thanks in advance and I apologize if I’ve missed something simple. I did notice the Linux image type, maybe that’s my answer?

Pardon if this thread is going out of date but I thought I’d chime in a bit since I’ve been struggling with this on a new install of FOG 1.2.0 on Ubuntu 12.04 and I “think” I’ve got my issue sorted a bit, at least enough that I had success multicasting this morning when it wasn’t working Friday afternoon. Note I’ve also had to figure out the tftpd-hpa restart thing which I don’t get at all. I’m an Ubuntu/Linux noob but I’ve got several FOG 0.29 boxes at different schools on Ubuntu 10.04 that have been running without problems out-of-the-box (so to speak) for YEARS so it’s unfortunate to see this kind of issue in an updated OS but it is what it is. Anyway, my clients were multicasting fine Friday morning on my new install until I shut it down to move it out of the lab and into our server room and a new switch too. All of a sudden I can’t multicast anymore but tasks have been going through this switch just fine all day so that ain’t it. Tons of Googling and forums research later and I get to this thread and a couple other like it so I vpn in from home (when I shouldn’t be working) and …

… from CLI I tried to restart the FOGMulticastManager and got this:

[EMAIL]administrator@GCC-FOGSERVER:~$[/EMAIL] sudo /etc/init.d/FOGMulticastManager restart

• Restarting FOG Computer Imaging Solution: FOGMulticastManager
  start-stop-daemon: warning: failed to kill 1069: No such process [ OK ]

---

… so thinking “AHA!! The service wasn’t even running” I tried several times to start the service and got this:

[EMAIL]administrator@GCC-FOGSERVER:~$[/EMAIL] sudo /etc/init.d/FOGMulticastManager start

• Starting FOG Computer Imaging Solution: FOGMulticastManager [fail]

---

… after thinking on it a bit I try to stop and start as individual steps:

[EMAIL]administrator@GCC-FOGSERVER:~$[/EMAIL] sudo /etc/init.d/FOGMulticastManager stop

• Stopping FOG Computer Imaging Solution: FOGMulticastManager [ OK ]
  [EMAIL]administrator@GCC-FOGSERVER:~$[/EMAIL] sudo /etc/init.d/FOGMulticastManager start
• Starting FOG Computer Imaging Solution: FOGMulticastManager [ OK ]

---

… Oooooooh, that looks good so I try another restart to see what happens:

[EMAIL]administrator@GCC-FOGSERVER:~$[/EMAIL] sudo /etc/init.d/FOGMulticastManager restart

• Restarting FOG Computer Imaging Solution: FOGMulticastManager [ OK ]

---

I was then able to successfully initiate a multicast task and the machines thankfully all woke up and did their thing!! BTW - I LOVE when I can see the progress of a multicast task in 1.2.0 which is the only way I could tell if anything worked since I was at home. It’s my favorite.

Something is goofy on a reboot where (in addition to tftpd-hpa issue) the FOGMulticastManager process can’t be started or restarted because “no such process”. I hope this may help someone who has the problem.

Thanks again for a quick reply! We do use the clone flag and when there’s no thawspace we have no issues at all. Guessing it’s a fog version thing we’re running into now since we are still at .29.

Wow, quick reply! Thanks!! Of course I’m now confused as I thought I had read on a number of occasions that fogprep was crucial for Win7 images to avoid boot errors. Or maybe that’s with earlier fog versions (we’re still using .29) since we have had some issues with .32.

Hoping someone is using Deep Freeze thaw spaces in Windows 7 successfully. In my case it’s not working out.

We use Deep Freeze with a small thaw space but when we run fogprep it’s evidently deleting the drive information in the registry and the thawspace is corrupted and unusable after deploying the new image.

Does anyone have ideas on how to get around this?

I had this issue on a new build and realized just now that I had put my server address into the gateway field in the network connection setup. When I put in the proper gateway and restarted it works like a charm.

I’m using 0.32 on an Ubuntu 12 server.

On some models of laptops when I upload a ~34GB image, they indicate that data copied will be 100+ GB and the image takes forever to upload, just as if a 100+ GB has been uploaded. However, when I then look at the actual image size on the server, it’s in the area of 30GB as one would expect.

On other models of laptops and desktops I have no problems whatsoever. Seems like the size of the hard drive has everything to do with it. The more available storage space, the larger the “Date Copied” shown in the progress screen and the longer the upload time. For example (all are the same image):

1. 80GB hard drive, 34GB “Data Copied” indicated and 30 min upload time.
2. 250GB hard drive, 100 GB “Data copied” indicated and 1hr+ upload time
3. 500GB hard drive, 202GB “Data Copied” indicated and 2hr+ upload time

Any ideas?

Also, when I try to use Single disk resizeable on my Windows 7 images, they appear to upload properly but won’t boot when deployed. I FogPrep them last thing before I shut them down to upload so I’m not sure why this is happening. I have no problem with Multipple Partition deployments except for the upload concerns above. I’ve looked through the Forums but found no-one having my particular flavor of problems.

Thank you for the response. Just in the last hour I seem to have hit on the right kernel ( 2.6.39.1 Core ) as my problem machine took an image. No knowing which kernel can be a challenge, but I can’t complain since you do get the option to name them. Thinking we were up-to-date I went backwards and nothing worked so I decided to start fresh and voila! Thank you for the reply!

Hello? Has no one seen this? I know it was commented on in the old forums, is there anyway to retrieve those comments?

I always get this when the image was created on and uploaded from a drive larger than the target drive you are now using. Maybe there’s a fix, but our workaround is to always build our images on drives we know will be smaller than the target.

FOG stops deployment task at “Checking queue” on client machine. Machine is a Dell 790, fog client 0.32 on an Ubuntu 10.04 release. Client just displays a count down (up) every 5 seconds. There were some references to this in the old forums which are now unavailable, hopefully someone can tell me how to resolve.