Issues with Downloading images FOG 1.2.0 CentOS 7
-
Well, I know FOG was working prior to setting up the firewall with certainty. I suspect it’s got something to do with the firewall, but I have no idea what it could be at this point. I’m wondering if for some reason the partclone software just isn’t being allowed to be called, because as I mentioned I can do the image manually using the debug download. Are there any known issues with 1.2.0 not being able to call partclone?
-
So I went back and did some testing, it does still work without the firewall enabled. I guess the next thing is to verify the ports I need opened to get FOG running, though I followed this: https://wiki.fogproject.org/wiki/index.php/FOGUserGuide#Full_Listing_of_Ports_used_by_FOG_server_and_client and I’ve gotten basically all those and then some enabled. Will let you know if I should find a fix or figure out the cause.
-
You could do a TCP Dump on the FOG server (without firewall) and run a unicast, and then a multicast task (just start them, see the partclone screen, see 1% reached, then turn those clients/tasks off).
Look through your dump, make note of every port used.
https://wiki.fogproject.org/wiki/index.php/TCPDump
And, if you figure this out, please let us know what ports you had to open.
-
Well I just tried to do that, and it’s failing to launch partclone again, this time without the firewall enabled. I can still call partclone manually, but it never seems to get called in the image script. I ran the Debug Download and it just says that it’s updating the database status, then says the task is completed. Partclone doesn’t even come up. This is absolutely bizarre, since this is the exact same machine setup that I was just able to clone from a little while ago.
-
@saistrop Try to reboot the FOG server, try again?
-
No dice, still doing the same thing. I can’t seem to pin down what’s causing the issue. Could it be that when I created the image, the machine had an SSD and an HDD, but the machine I’m attempting to image only has an SSD?
-
Can you post a photo (or video) of exactly what you’re seeing?
Or give us the exact errors you’re seeing? A photo is much preferred.
-
Will do when I’m able, but in the mean time I’m attempting to see if an upgrade to the git branch will help. Do you guys have any suggestions for installing the current branch on CentOS 7? I’m getting an issue with mysqldnd missing during install.
-
The Trunk versions come from sourceforge, and they have been having… issues… lately.
I recommend using this method for upgrading to FOG Trunk:
https://wiki.fogproject.org/wiki/index.php/SVN -
Would you recommend using SVN or Git for the current trunk? I tried with Git, and got to “Downloading current FOG client” and then it would error out. Also, I’ve attached a picture of the issues I was having using the Debug Download, partclone never seems to get called and it goes straight to task completion. Pictures are from most recent to least recent.
-
This is a picture of some of the errors I’m getting while installing the Trunk version. I commented out the exit status 1 for the configuring MySQL function like recommended in the wiki, but I get an error when downloading the FOG client:
-
When I get home, I’ll download the current Trunk version and just stick it into my MEGA account for everyone to use.
-
@Wayne-Workman I would certainly appreciate it, I know it’d likely be easier if I rolled back to CentOS 6.6 but my company is trying to push everything up to the most modern version, and they don’t always realize new isn’t always better.
-
Try to install mariadb-server first…
[CODE]yum install -y mariadb-server[/CODE]
and to disable exiting if FOG Client download fails, edit this:
trunk/lib/redhat/functions.sh
Find the code block that looks like this (in vi, the forward slash searches for strings…):
[CODE]echo -n " * Downloading New FOG Client file…";
cwd=pwd;
cd $webdirdest/service;
count=0;
while [ -z “$clientVer” -a “$count” -le 10 ]; do
clientVer=wget http://$ipaddress/fog/service/getclient.php -q -O -;
if [ -z “$clientVer” ]; then
clientVer=wget http://$ipaddress/service/getclient.php -q -O -;
fi
count=expr $count '+' 1
sleep 2;
done
if [ -z “$clientVer” ]; then
echo “Failed to get client version”
exit 1
fi[/CODE]And put a hash before the exit 1 like this, then save:
[CODE]echo -n " * Downloading New FOG Client file…";
cwd=pwd;
cd $webdirdest/service;
count=0;
while [ -z “$clientVer” -a “$count” -le 10 ]; do
clientVer=wget http://$ipaddress/fog/service/getclient.php -q -O -;
if [ -z “$clientVer” ]; then
clientVer=wget http://$ipaddress/service/getclient.php -q -O -;
fi
count=expr $count '+' 1
sleep 2;
done
if [ -z “$clientVer” ]; then
echo “Failed to get client version”
#exit 1
fi[/CODE]But please note that if the client download fails, that means you DO NOT have the new client.
You can always run the installer later (no harm) and try to get it.
-
MariaDB is already installed and updated, would it be worth trying to remove it and let FOG install it? And that got me farther, now it’s failing on the Plugins. Unfortunately it’s about closing time here so I’ll have to get back to this on Monday. I really appreciate the help, hopefully a new stable build is coming soon that’ll work better with CentOS 7’s weird setup.
-
I ran the FOG installer and got the exact same error about MySQL…
I installed mariadb-server and then re-ran the installer and the error went away.
-
That’s strange, since I’ve already got MariaDB installed. Maybe I’ll try to remove it and install it again, though I don’t think that would have much effect at this point. I still haven’t found a good resolution for the issue, at this point I may just end up having to start FOG over from scratch, which would be unfortunate, but it may fix it.
-
@saistrop said:
good resolution for the issue, at this point I may just end up having to start FOG over from scratch, which would be unfortunate, but it may fix it.
FYI, I installed CentOS 7 at the house and gave Tom access, and he fixed up the current FOG Trunk to work on it…
If you update FOG and try again, there’s probably a high likelihood that it’ll work.
-
Well, minor update. I got the current SVN version from updating (3526 is what’s listed), and it seems to be having an issue with the old image I had from 1.2.0. Is there a known way to work with old images on the current SVN?
-
Yes.
First thing is to verify that the path is correct. Nevermind the image name, look at general - > path.
This must exactly match the actual path of the image, and it is case sensitive.
Second thing is to make sure FTP is good, third thing is to make permissions 777 recursively on the images directory.
and if the image was made with 1.2.0, it would be using partclone (not partimage).
