Swapin problems with User custom Image and disk geometry???

[burnettb317]

Sorry, I know Friday afternoon is a terrible time to post a question, but I thought I'd get it out there for Monday morning :)

We have been using part of our Emulab as build nodes for a DevOps autobuild system.  At the moment the image being used is a user image, and isn't exactly optimized, it is based on the UBUNTU18-04 image from you and we've added docker to it:   UBUNTU18-04-DOCKER.

The user has been seeing a lot of problems swapping and I just heard about it taking a long time to swap since it is only successful 1 our of 5 attempts and it runs out of machines while they are reloading so it fails , and eventually it works.  So I did a test swap in of 2 different machine types and watched the console.

The nodes play frisbee fine and the image is transferred, it is rather large, original is about 700MB and the new image is 1.4GB, but the image transfer completes, I checked the logs.  then when the nodes reboot there are issues with the console type and geometry and It's been awhile since I've looked at these logs so I think there are multiple errors, but I forget how to read these and what a lot of it means.  I've included an excerpt of the swap log right after the frisbee and during the boot until the fail/time out and boot back to the mfs, Looking for some help deciphering what is wrong.  

Thanks for the Help

-Ben

Log from Machine Type 1 pc64 (Dell_E6550):

Left the team after 517 seconds on the field!

Set partition 1 type to 0x0083

Wrote 17179869184 bytes (4230434816 actual)

0 8314 24967

Fri Jan 24 14:26:43 CST 2020: Image #0 load complete

Fri Jan 24 14:26:43 CST 2020: Frisbee run(s) finished

Fri Jan 24 14:26:43 CST 2020: Running slicefix

Fri Jan 24 14:26:43 CST 2020: Adjusting slice-related files on ad8 slice 1

WARNING: console type not set, assuming "sio1"

*** ad8s1:

slicefix: linux: unknown disk ad8

Fri Jan 24 14:26:43 CST 2020: WARNING: slicefix on ad8 slice 1 failed

Fri Jan 24 14:26:43 CST 2020: slicefix run(s) done

Fri Jan 24 14:26:43 CST 2020: Resizing final disk partition

growdisk: using DIOCG* to determine size/geometry

/dev/ad8: no change made

Fri Jan 24 14:26:43 CST 2020: Waiting for server to reboot us ...

IPOD: reboot forced by c0a80a03...

cpu_reset: Stopping other CPUs

Consoles: serial port

BIOS drive A: is disk0

BIOS drive C: is disk1

PXE version 2.1, real mode entry point @9b85:0106 (!PXE)

Log from Machine Type 2 pc55 (HPdc7700):

Left the team after 498 seconds on the field!

Set partition 1 type to 0x0083

Wrote 17179869184 bytes (4230434816 actual)

0 9876 24967

Fri Jan 24 14:26:43 CST 2020: Image #0 load complete

Fri Jan 24 14:26:43 CST 2020: Frisbee run(s) finished

Fri Jan 24 14:26:43 CST 2020: Running slicefix

Fri Jan 24 14:26:43 CST 2020: Adjusting slice-related files on ad0 slice 1

WARNING: console type not set, assuming "sio1"

*** ad0s1:

  disabling time-dependent fsck

  fixing Linux root partition ad0s1

  updating /etc/fstab

  updating /boot/grub/grub.cfg

    setting grub2 root to hd0,1

    setting console to sio1

  localizing ...

  updating /root/.ssh/authorized_keys

  updating /etc/localtime

cp: /mnt/etc/localtime: No such file or directory

Failed to create /etc/localtime

Fri Jan 24 14:26:44 CST 2020: WARNING: slicefix on ad0 slice 1 failed

Fri Jan 24 14:26:44 CST 2020: slicefix run(s) done

Fri Jan 24 14:26:44 CST 2020: Resizing final disk partition

growdisk: using DIOCG* to determine size/geometry

/dev/ad0: defining partition 4 as start=46139392, size=110162096

/dev/ad0: partition table modified

Fri Jan 24IPOD: reboot forced by c0a80a03...

 14:26:44 CST 2020: Waiting for server to reboot us ...

cpu_reset: Stopping other CPUs

Consoles: serial port

BIOS drive C: is disk0

PXE version 2.1, real mode entry point @9bca:0106 (!PXE)