Latest FOG 0.33b
-
I am a little confused,
[quote=“fabritrento, post: 22429, member: 21607”]
(partclone gray screen):
/dev/sda1 is mounted at …
error!/dev/sda2 is mounted at …
error!/dev/sda3 is mounted at …
error!
[/quote]There’s no more detail than just error?
[quote=“fabritrento, post: 22429, member: 21607”]
after this first fail of download
[/quote]
What do you mean by after the first download, weren’t you just doing a Wipe Task?[quote=“fabritrento, post: 22429, member: 21607”]
if i restart the same image download with a new download command on the web gui (same pc) the download starts… and finish but the grub installation don’t work, i must chroot the pc and do grub-install /dev/sda[/quote]Download jobs of the system, particularly pertaining to Grub, have never seemed to work. I’m currently trying with count=62 rather than count=63 on the mbr recovery. This probably means you’ll need to reupload the job first to see if it works. It may even be simpler than this. That said, I don’t think grub and mbr together worked properly and we haven’t come up with a good fix to this.
While I can understand the need to image such systems with multipartition, my goals for right now are to get the essentials (the things more people use) working properly. Windows 7, Windows XP, Windows 8/8.1, and Linux. Not multiboot systems. While I am trying to get this working, it’s also set me behind in getting other areas actually working.
Right now I’m trying to figure out why I need to use parted after reloading the mbr/partition table. The partition sizes should be created, with the only exception being on resizable images. Resizable images should need the partitions directly told as we want to be able to use the full size of the disk. MPS/MPA should not need to have their partition tables re-established as we recover the original partition table. Just my thoughts.
I am not saying I won’t try to get this fixed, I’m just saying I’m busy on other areas first.
-
Hi Tom,
there is the “ls” of the image folder:
-rwxrwxrwx 1 root root 32K feb 3 23:30 d1.mbr
-rwxrwxrwx 1 root root 219K feb 3 23:30 d1p1.img (ntfs empty)
-rwxrwxrwx 1 root root 1,2G feb 3 23:34 d1p2.img (/ of ubuntu)
-rwxrwxrwx 1 root root 69 feb 3 23:34 d1p3.img (extended)
-rwxrwxrwx 1 root root 2,3M feb 3 23:36 d1p5.img (/home ubuntu)
-rwxrwxrwx 1 root root 3,2M feb 3 23:35 d1p6.img (swap)All partitions are present.
-
r1190 released.
I’ve tested multipart upload and download, both seem to work perfectly fine. I think joacliba’s issue is due to the “extended” partition as the files are looking for partition numbers, not logical partitions. I don’t have a good method for this, maybe use RAW for this specific type of image.
You can now search for tasks on the task management page. This means, completed tasks, canceled, hostname task type etc… many different choices for search.
Thank You,
-
It’s really weird. I’ll test more tomorrow, I have another theory for what is failing.
Thank you Tom!
-
r1191 released to fix header row alignment issues. Adds a few other changes as well.
-
r1192 released.
Adds the r8106e drivers to the kernel.
Adds checking if the drive is gpt or mbr on a very simplistic method. It uses gdisk -l /dev/sda ($hd for dynamics) and greps for the GPT areas. If gpt is not present, it does mbr backup methods, if it’s anything else (present usually) it backups up the GUID Partitioning tables (GPT). This is only needed during the upload of the image for mps or mpa (multi-part single/multi-part all)
On the download, it checks the size of the mbr file created (yes on upload it still uses the .mbr identifier for simplicity sake). If the size is 512, it’s not linux but loads as mbr methods do. if it’s 32256, it is linux mbr and loads as such. If it’s anything else, it assumes it’s GPT as I don’t imagine GPT gets any larger or smaller anyway.
Like I said, very simplistic but should do what we’ve kind of been stuck on for a little bit.
-
Tom,
Are there still problems with deplopying a windows 7 image with image type “Multiple Partition Image - Single Disk”?
I deployed a uploaded window 7 image with 2 partitons (system reserved and system partition) to a pc with a clean hard disk. Partclone exits at the first time with only the message “error exit”, after I started the task again, partlcone succesfully deployed the image to the pc.I used revision 1192.
Ty,
Lukas -
Does it specify the error at all?
-
r1193 released.
Should properly create GPT tables, but I don’t know as I don’t have GPT disks to test, so please let me know.
Fixed a time sending issue on autologout service.
-
[quote=“Tom Elliott, post: 22519, member: 7271”]Does it specify the error at all?[/quote]
Partclone shows only “error exit”, but directly after partclone exits, I saw that “error: /dev/sda1 no such file or directory”. I booted then into debug mode with fog and executed “fdisk -l”. It lists me /dev/sda1 with system HPFS/NTFS/exFat.
Looks like that the partitions are not there at the moment, where partlcone tries to deploy the image.Edit:
Forgot to mention that at upload, fog shows me something with “unary operator expected” in file fog.upload at line 258, 262, 266. -
I’m saw the error, but the MBR data is correct and get’s written as we all of a sudden see /dev/sda2.
Are you sure it said “error: /dev/sda1 no such file or directory” and not “error: /dev/sda2 no such file or directory”?
-
I tested it again and maked a screenshoot and there I see /dev/sda1. How I wrote in my post number 830, something goes wrong at the first time, but after starting the deploy again. Partclone and fog then works fine.
-
Alright,
Found the issue. Needed to wipe out the MBR/GPT structures which is now implemented.
r1194 released should correct this issue.Thank you,
-
[quote=“Tom Elliott, post: 22525, member: 7271”]Alright,
Found the issue. Needed to wipe out the MBR/GPT structures which is now implemented.
r1194 released should correct this issue.Thank you,[/quote]
I updated to new revision, same problem. After starting the task again, all works fine then.
-
[quote=“Albatros, post: 22521, member: 16710”]Edit:
Forgot to mention that at upload, fog shows me something with “unary operator expected” in file fog.upload at line 258, 262, 266.[/quote]I found this issue and already fixed it. It’s because of the variable call in the script. Sorry I forgot to wrap them in double quotes.
r1195 released should fix that issue.
-
[quote=“Albatros, post: 22526, member: 16710”]I updated to new revision, same problem. After starting the task again, all works fine then.[/quote]
Did you restart the system after updating. If not, it’s still using the old/bad init.gz
-
I updated to revision 1195,restarted the system and overwrote the init.gz with that from the subversion directory.
Sorry same problem. -
Can you do me a favor Albatros,
I’ve seen on a couple instances where the installfog.sh doesn’t update all the files appropriately, for some reason.
Can you try directly copying my init.gz to your /tftpboot/fog/images folder, reboot the client and try again?
-
I deleted the init.gz from /tftpboot/fog/images/ directory and placed there your init.gz. Same problem .
I am now away, see you later -
If you get a chance, can you try re-uploading the image, then deploy?
I’m thinking the MBR checking got screwed up somewhere and this is why it’d doing a double deploy issue. I’m not seeing the issue anymore.