Deploy image not expanding partitions
-
@Tom-Elliott said in Deploy image not expanding partitions:
k -fp /dev/sd1
Hello, I have the result :
-
@Tom-Elliott The new picture have all results of test all partitions.
-
@SuperMajestic the sda3 didn’t need to be done, only ext formatted volumes need to be tested.
-
for clarity, this was done before FOG ran, but after clean install, right?
-
@Tom-Elliott Yes, sorry, sda3 is a swap partition.
-
Yes.
- Without fog, I installed fresh installation of Centos 7. I see that it works. The installation is finished correctly, and I see the root login.
- I created the host on fog client and created the capture task.
- Reboot the server with force pxe.
- were Typed e2fsck -fp /dev/sda1
- I’m waiting your instructions before run “fog” command.
-
@Tom-Elliott More info, I print fdisk result:
/dev/sda1 is a necessary partition to bios with GPT, type “BIOS boot”.
/dev/sda2 is a boot partition
/dev/sda3 swap
/dev/sda4 /
-
This is the last test. Now fsck is correct but crash in resizing sda2 (/boot).
-
This is the content of generated d1.partitions
-
@Tom-Elliott it’s incredible
Now can’t capture nothing. gpt and mbr not works. Nothing :-S. Error in resizing sda1 even in two different hosts. -
@SuperMajestic The capture logic hasn’t changed. Would you mind re-running the installer though?
Another dev found that there were two issues accidentally added to the capture. (But have been fixed for a bit).
The start position during our resizing was accidentally being shifted around (which could cause what you’re describing). I’ve tested capturing both gpt and mbr, and all is working from my side. I’ve tested deploying both gpt and mbr now and am fairly sure I’ve fixed the issues with the resizing now too.
So please, re-run the installer and hopefully things will go better.
-
Hello, I downgrade to stable version (Running Version 1.3.4 SVN Revision: 6066) and don’t capture.
The error isn’t in resizing, is in checking.Do you want I upgrade fog to latest version again?
-
@SuperMajestic Yes please.
I believe the issue was the start point was trying to be shifted around.
The resizing, at least for linux systems, has to try to get the data first. We have to see how far it can be shrunk.
