Casper Imaging stalling after a few batches

fsuarez
New Contributor

Okay so i want to see if anyone is or has experienced this besides me. We have our jss and 3 distribution points which are running mountain lion. We can image 8 to ten computer at a time but after we do two or three batches that number decreases and images start to fail. We then need to log on to the server and clear out the client0 folder restart the server once sometimes more and can then image again.

3 ACCEPTED SOLUTIONS

CasperSally
Valued Contributor II

Before you clean out the folder, is the server running out of space?

View solution in original post

donna_mckee
New Contributor

I am having the same problem either the image stalls out or fails. we are not out of space and the not using full Ram to it fullest .

View solution in original post

fsuarez
New Contributor

@CasperSally Okay checked the server and it is not out of space. We are noticing that the ram on the server is almost all in use but mostly with inactive processes. It shows that 7.6 gigs are being used and when we restart it it shoot up to less then one in use.

View solution in original post

14 REPLIES 14

CasperSally
Valued Contributor II

Before you clean out the folder, is the server running out of space?

donna_mckee
New Contributor

I am having the same problem either the image stalls out or fails. we are not out of space and the not using full Ram to it fullest .

fsuarez
New Contributor

@CasperSally Okay checked the server and it is not out of space. We are noticing that the ram on the server is almost all in use but mostly with inactive processes. It shows that 7.6 gigs are being used and when we restart it it shoot up to less then one in use.

oopsydaisy
New Contributor III

I am having same issue. From what Jamf has told me, it is a problem with Apple's diskless implementation in NetBoot. This seems to make sense as I didn't have this issue with DeployStudio, which does not use a diskless method, if I remember correctly. Still sucks, though

CasperSally
Valued Contributor II

Have you noticed the shadow files getting very large, even if they aren't filling up your drive (yet)?

If so, check the solution Kumarasinghe posted here. We're about to start mass summer imaging here, but our techs image a few dozen computers a day and I haven't seen this issue here (I use the solution posted there to keep shadow files small).

https://jamfnation.jamfsoftware.com/discussion.html?id=6443

oopsydaisy
New Contributor III

I've tried this fix and it hasn't helped. My shadow files are between 200-300MB and there's about 400GB available on the drive. The most I have ever been able to image was 6 before I had to delete the shadow files, and most times, reboot the server. I really hope V9 helps with this issue somehow. Does anyone think upping the actual RAM would help with this?

CasperSally
Valued Contributor II

Strange, I wonder what the difference is between our environment, we're 8.64 but I haven't had to reboot our netboot servers in months. We image probably hundreds a week off of 4 servers.

oopsydaisy
New Contributor III

Are you using external storage? What version OSX are you using?

CasperSally
Valued Contributor II

We are using 128GB SSD drives in XServes for netboot. The XServes are 10.7.4, and we are using 10.8.3 Netboot images. All of our diskless problems went away when we modified our netboot rc.netboot file. Our shadow files are only 50-150MB.

Right now techs prob average imaging 5 at a time per server, but they image day after without any server issues. Knock on wood as we'll be mass imaging thousands of machines over next few weeks.

oopsydaisy
New Contributor III

Maybe it's an issue with 10.8 NetBoot service as my DeployStudio server was running 10.7.4 and didn't have an issue. But, as I was reminded by Jamf, DeployStudio doesn't use diskless, whereas Casper does. I'd be curious to know if all the people having this issue are running 10.8 servers. I'm running 10.8.2, just for clarification

CasperSally
Valued Contributor II

I sure hope the JAMF 9.0 JDS servers make all these netboot headaches a distant memory.

oopsydaisy
New Contributor III

What she said:)

hkim
Contributor II

Diskless is basically borked in 10.8 for one reason or another from my experience. It works half the time at best.

Change your netboot sets to use a RAM disk instead for shadow files.

http://www.macos.utah.edu/documentation/administration/setup_netboot_service_on_mac_os_x_10.6.x_client.x_client/setup_netboot_service_on_mac_os_x_10.6.x_client-diskless_netboot.html

These are not exact instructions but guidelines on what to do and editing the rc.netboot file to get the results that you are looking for.

alexjdale
Valued Contributor III

I ended up dropping Netboot this year and it has been a great decision. In my opinion, Netboot is poorly supported by Apple and is prone to various problems which can be difficult to troubleshoot.

I moved to USB boot media only, and this allowed us to decomm ~40 Mac Mini Servers and has significantly reduced support escalations. Booting is faster and more stable and I can just publish the DMG to Box for our technicians to access whenever they need it.

We also had issues with TFTP being flagged as a risk by our security team and we are implementing wired 802.1x for all networks, so the writing was on the wall anyway.