Latest FOG 0.33b
-
Is the snapin actually running?
I see many different items.
Can you show me log of /var/log/apache2/error.log or /var/log/httpd/error_log
-
That’s the issue if you look its all the same snapin I launched it once and for some reason its duplicating the task it is running but it is running muiltiple times and not just once. attached is the error.log from apache2 thanks for looking into it.
[url=“/_imported_xf_attachments/0/502_error.txt?:”]error.txt[/url]
-
r1119 released.
Please give this a try. I don’t know why it keeps generating a new snapin-task for you. I’ve not added that to the Task mangement page. I think i found the issue though. The way the class was being called it was actually saving which would create that particular issue.
Just run:
TRUNCATE table snapinTasks;
TRUNCATE table snapinJobs;Then retry to send the task.
-
r1120 released to hopefully fix the same error for active-multitasks during the state lookup.
-
Hi Tom
i make some modification on fog script to support gpt partition
for resize partition, all partition and all disk
resize with gpt was working but for now don’t take in consideration all the case
all part and all disk don’t work for now,
i put the file only for you to see them.
do you think you can use them (after debug)[url=“/_imported_xf_attachments/0/505_fog.zip?:”]fog.zip[/url]
-
the issue is still happing I tried a different snapin also just to ensure it was not the application it self. attached is the screen shot of the page do you need the error.log again?
[url=“/_imported_xf_attachments/0/506_fogchromeerror.pdf?:”]fogchromeerror.pdf[/url]
-
I’m trying to replicate this issue, but so far I’m getting nothing.
-
do you think dropping the database and reinstalling fog may help? I would have to start everything all over but it may be worth it if it fixes the issue what kind of snapins are you trying. it only happens after it deploys on the ws.
-
Just a feature request because I am rebuilding everything from scratch now… Is there a way to be able to export the snapin part of the data base or all of your fog settings? I know you can with the inventory. but I think it would be nice for people who have more then 10 snapins if they ever had to rebuild.
-
i debug the file ,
multi part, and all part work now with different partition type
i will test tomorrow gpt part[url=“/_imported_xf_attachments/0/507_fog.zip?:”]fog.zip[/url]
-
I’m currently rewriting the Service scripts that deals with handling the snapin tasks. I’m in the midst of testing. I think it was creating a new entry when it wasn’t really necessary. I could set all the task/snapin task variables (theoretically) by the service scripts them selves, rather than having to use the ClientServiceManager.class.php file. It seems like this is the culprit to your particular issue Troye. It means, however, that things are neater and more in place. I can actually, hopefully, remove ClientServiceManager altogether and still get things to work. Less code, easier to troubleshoot and maintain. It does, however, mean reenabling the Task creation part of the snapin deploy job, though I imagine it will be easier to cancel the task by the active tasks. I just need a way of making sure it doesn’t fail out because you’ve set a deploy to the job, then want to upload the image when you’re done deploying. That’s for the future though.
-
[quote=“Tom Elliott, post: 21659, member: 7271”]I’m currently rewriting the Service scripts that deals with handling the snapin tasks. I’m in the midst of testing. I think it was creating a new entry when it wasn’t really necessary. I could set all the task/snapin task variables (theoretically) by the service scripts them selves, rather than having to use the ClientServiceManager.class.php file. It seems like this is the culprit to your particular issue Troye. It means, however, that things are neater and more in place. I can actually, hopefully, remove ClientServiceManager altogether and still get things to work. Less code, easier to troubleshoot and maintain. It does, however, mean reenabling the Task creation part of the snapin deploy job, though I imagine it will be easier to cancel the task by the active tasks. I just need a way of making sure it doesn’t fail out because you’ve set a deploy to the job, then want to upload the image when you’re done deploying. That’s for the future though.[/quote]
This sounds promising can’t wait until its ready for me to test in the mean while I am rebuilding another server to see if I can recreate the current issue also.
-
After much tribulation, I think I’ve finally gotten snapin tasks working.
It’s performed by the Host class (meaning I can get rid of the two functions creating the packages. within the ClientServiceManager).
Hopefully you’ll see it working, I don’t have a true to form, yet, snapin to test directly.
-
Ok just let me know when I can try the new svn
-
You can try it now. I’m still refining a little bit though.
-
[quote=“Tom Elliott, post: 21627, member: 7271”]No, i mean from the image page.
You can’t assign two OS’s to the image name. If you’re trying to image a dual boot, you need to setup RAW image type versus MPS/MPA/SDR[/quote]
this is not correct, because with 0.32 version, i already done a dualboot (sda1 winxp sda2 linux) imaging upload and then deploy without problems setting OS to windowsXP
i tryed single disk multiple partition image upload with dualboot win+linux, and works in 0.32. In 0.33b don’t work (to be fixed)
i’m now trying multiple disk multiple partitions in 0.33b, then i post here the result.on 0.32 all working, except the grub 2.0 that was corrupted after imaging process (i don’t know how)… but is not a big problem, is sufficient to boot client in debug mode (via grub fog menu) and do:
mount /dev/sda2 /mnt
mount -o bind /dev /mnt/dev
mount -o bind /proc /mnt/proc
mount -o bind /sys /mnt/sys
chroot /mnt
update-grub
grub-install /dev/sdai think that the best is to add this as post deploy script on clients with ext fs (generally means that linux is present on machines…)
it can be simply checked doing a fdisk -l|grep ext then take the partition string sdX and substitute to mount and chroot and grub command.
additionally we can check if /boot dir is present in the extfs (to be sure that is a os filesystem).I see that on 0.33 the /opt/fog/log dir is empty.
and 2 process (fogscheduler and fogmulticast) is not running ad deamons, is this correct?
-
i’m trying to debug the problem that imaging of single disk, multiple partition don’t work.
so i started a debug upload.the screen say to me to type fog to start image upload.
here images uploaded to server:
administrator@fog:/images/labciro$ ls -la
totale 6890908
drwxrwxrwx 2 root root 4096 gen 21 08:54 .
drwxrwxrwx 4 root root 4096 gen 21 08:54 …
-rwxrwxrwx 1 root root 512 gen 21 08:35 d1.mbr
-rwxrwxrwx 1 root root 7056262600 gen 21 08:54 d1p1.img
-rwxrwxrwx 1 root root 20 gen 21 08:54 d1p2.img
-rwxrwxrwx 1 root root 20 gen 21 08:54 d1p3.img
administrator@fog:/images/labciro$gunzipping d1p2.img and d1p3.img the result is empty file. so seems that the image upload is not started…
partitions on pc are:
sda1 30GB ntfs winxp
sda2 15GB ext4 xubuntu
sda3 1GB linux-swapso sda2 and sda3 is not imaged (only 20byte is copied).
here the value of variables:
##########
osid=1
osname=Windows XP
mbrfile=/usr/share/fog/mbr/xp.mbr
type=up
storage=192.168.0.3:/images/dev/
img=labciro
hd=/dev/sda
part=/dev/sda1
##########i think that in these variables the info that also sda2 and sda3 is to be done is missing. do you confirm? how i can force to do image of: sda1,sda2,sda3 and not sda1 only??
where i can find the fog script on the server? in what directory is stored? (to debug)
i tryed to run fogpartinfo script and detects 3 partitions (right) so i must understand why only first 20byte is saved on server of 2nd and 3th partitions…
now i’m running fog script by hand, after that i’ll check the log on the client:
the problem with second and third partition is from :
ntfsclone-ng.c: no ntfs partition, ntfs mount error 22i think that partclone is launched with bad options…
i think that we must modify fogpartinfo script to detect list of partitions and also the partition type (fdisk -l output is enough) so we can start the differents partclone commands in right way.
here contenents of client log:
administrator@fog:~/test_fog$ more fdisk-after
Disk /dev/sda: 160.0 GB, 160041885696 bytes
255 heads, 63 sectors/track, 19457 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytesDevice Boot Start End Blocks Id System
/dev/sda1 * 1 3824 30716248+ 7 HPFS/NTFS
/dev/sda2 3825 5648 14647296 83 Linux
/dev/sda3 5648 5773 1000448 82 Linux swap
administrator@fog:~/test_fog$
administrator@fog:~/test_fog$ more fdisk-beforeDisk /dev/sda: 160.0 GB, 160041885696 bytes
255 heads, 63 sectors/track, 19457 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytesDevice Boot Start End Blocks Id System
/dev/sda1 * 1 3824 30716248+ 7 HPFS/NTFS
/dev/sda2 3825 5648 14647296 83 Linux
/dev/sda3 5648 5773 1000448 82 Linux swap
administrator@fog:~/test_fog$
administrator@fog:~/test_fog$ more partclone.log
Partclone v0.2.69 [url]http://partclone.org[/url]
Starting to clone device (/dev/sda3) to image (/tmp/pigz1)
Reading Super Block
ntfsclone-ng.c: NOT NTFS partition, ntfs mount error 22
administrator@fog:~/test_fog$we must do modify fog script to check the partition/fs type and adapt the partclone command in various situations.
-
Hi, I’m not sure if you know about it (search didn’t found anything), but probably found bug in /var/www/fog/lib/fog/Host.class.php, line 850
There is a typo in [I]$MulticasetSessionAssoc->save();[/I] - should be without “e” in Multicaset, as after the change creating multicast task works - before I was getting blank screen.
Btw, can you say, how far is 0.33 from final release? (I asked you this in the other topic, but you probably missed it). Thanks
edit: I’m still using 1118, so perhaps it’s already fixed…
How is it possible to update between revisions? Just download from svn and bin/installfog.sh? Thanks -
I think Tom Elliott is busy today!
Tom, i need you! if you come in italy i can offer to you a gallon of beer!
And if fogproject needs hardware/support to work i can give help
-
i think that the error is in this part of fog script.
the function: determineOS $osid
is used to say to partclone what kind of partition we must clone.
this is not correct.
in the script:
if [ “$osid” == “50” ]; then
fstype=“extfs”
else
fstype=“ntfs”causes all partition that is not ntfs in a pc with windows system fail to image.
i have on disk:
sda1 ntfs (clones correctly)
sda2 ext4 ( is cloned as ntfs, is not correct)
sda3 linux-swap ( is cloned as ntfs, is not correct)we must modify the script to check via fdisk -l what kind of partition is really present
or more best:
sudo blkid /dev/sda1
/dev/sda1: UUID=“b842abee-3fbf-4347-a451-ec9e0b3b3aa8” TYPE=“ext2”We cannot know what kind of partition is present simply assuming by $osid, i can have a windowsxp in first partition as ntfs, ok, but in all others i can have a very differend kind of filesystem type!
so Tom, if you need help to test/develope ask me!
