Cloudmin 6.7 install removed my cron job

Hi,

Just updated 2 hosts with Cloudmin 6.7 using yum (CentOS 6 and CentOS 5), in both cases the installer removed my manually created cron job (using Webmin module) and put in Webmincron script, which broke the server because the script in question needs to be executed every minute, and not every 5 like Webmincron. Script name is "bandwidth_calc.sh"

Is there a module to edit webmincron jobs?

Thanks.

PS. Latest Virtualmin release is missing change log. Does it fix the multiple backup destinations bug?

Status: 
Closed (fixed)

Comments

That's odd, because only Cloudmin's own cron jobs should be converted to webmincron.

What was the exact script that your cron job ran?

tpnsolutions's picture
Submitted by tpnsolutions on Thu, 01/24/2013 - 13:10

Hi,

Change Logs can be found at:

Usermin Change Log: http://www.webmin.com/uchanges.html

Virtualmin Change Log: http://www.webmin.com/vchanges.html

Webmin Change Log: http://www.webmin.com/changes.html

*** these are generally published on the Webmin.com site days before they are released to Virtualmin.com as things go through some extra testing, however there are rarely an extra changes made to the log after release ***

-Peter

Exact cron job command is: sh /etc/windsolutions/bandwidth_calc.sh

I just tested creating a cron job like that manually and then re-running the Cloudmin installer, but it didn't modify the job (as expected, as it isn't a script under /etc/webmin/server-manager).

If you look in the /etc/webmin/webmincron/crons directory, is there a file that contains bandwidth_calc.sh ? And if so, what does it contain?

Yes there is

arg0=
mins=*
user=root
id=13590412654022
line=0
module=server-manager
arg1=/etc/windsolutions/bandwidth_calc.sh
hours=*
active=1
days=*
index=5
nolog=
func=run_cron_script
weekdays=*
type=0
months=*

That is really odd - clearly that conversion is wrong, but I can't see why Cloudmin would have done it.

Are you running Cloudmin Pro, or one of the GPL versions?

As a quick fix, I would suggest deleting that file in the /etc/webmin/webmincron/crons directory that contains /etc/windsolutions/bandwidth_calc.sh , and then re-creating your regular cron job.

I am still looking into the underlying cause of this bug..

I can confirm that on my Cloudmin Pro server that upgrading also removed my cron job. (which was an rsync job) My cron job looked like this before it was removed:

20 1 * * * rsync --remove-source-files -e ssh --progress --partial -avz --log-file=/rsynclog/rsynclog.log /mnt/backups/ root@216.108.146.33:/c/backup/virtual/vmhost

EDIT: Im on Ubuntu 12.04.1 LTS (GNU/Linux 3.2.0-36-generic x86_64) and updated using the Cloudmin GUI.

Franco - was it just this one cron job, or did others get effected too?

this server is dedicated to cloudmin, so there were no other cron jobs to speak of however I did note something that may help.

A fresh install of ubuntu usually fills the cron file with instructions in comments. these I usually ignore but leave intact in case someone who does not know linux needs to clean up my mess if I am not around. These instructions were there before the upgrade but not after, after the upgrade the instructions and my rsync cron job were missing and the only thing in my cron file was....

43 16 * * * /etc/webmin/server-manager/space.pl
@daily /etc/webmin/server-manager/backup.pl --id 1352938572112550

@daily /etc/webmin/server-manager/backup.pl --id 1352996120324840
0 * * * * /etc/webmin/bandwidth/rotate.pl
@daily /etc/webmin/server-manager/backup.pl --id 1358378510174360

I then re-added my single cron job Hope this helps... Franco

I have found the bug that causes this, and will be releasing an update shortly to prevent it in future..

Thanks for your bug report!

Thanks Jamie. It's v6.8 I suppose

Actually, haven't yet updated to the latest Virtualmin because of that issue. Does Vitualmin has the same problem and if yes, was it corrected?

Another issue, probably related, in version 6.7, it has not detected server status change of all managed servers, more precisely it has not detected that updates (Centos 5/6 yum) are available, or in one case already applied. After refreshing the status manually the updates were detected.

One more symptoms pointing to that is the successful SSH logins have dropped from few hundreds a day to zero.

Seems that v6.8 resolves the server status refresh.

Virtualmin 3.98 doesn't have this bug.

Cloudmin 6.8 fixes the lack of status refreshes, which is related to the original bug.

Sorry, but I just discovered that my low disk space alerts are not automatically checking either, is this also related or should I open a new bug?
Franco

Franco - where are those low disk space alerts setup? In Cloudmin, or elsewhere?

Yes they are in cloudmin, and by coincidence after reading about this bug, I went to add a disk space monitor, and clicked refresh and a whole bunch other disk space alerts turned on for my VMs so I thought I would ask if this was related...... F

Automatically closed -- issue fixed for 2 weeks with no activity.

Jamie The system auto closed this, but I want to know if its fixed or not.

The original issue, where updated Cloudmin removed a cron job, was fixed on Sun, 2013-01-27 (in comment #12 above). It is part of Cloudmin 6.8:

https://virtualmin.com/node/25107

Jamie would know best, but it sounds like what you're describing about a number of disk alerts turning on is unrelated to the above.

However, you could try upgrading to 6.8, and see if you're able to reproduce that issue again.

Version 6.8 fixed the original bug of Cron jobs being deleted.