GFI Software

Welcome to the GFI Software community forum! For support please open a ticket from https://support.gfi.com.

Home » GFI User Forums » Kerio Connect » I think i've found a backup issue/bug...
I think i've found a backup issue/bug... [message #74973] Thu, 28 October 2010 00:07 Go to next message
lythaby is currently offline  lythaby
Messages: 5
Registered: June 2007
I've just had to ssh in from home(Somewhat spoiling my bottle of Rosemount Shiraz), to delete the Kerio Backup Folder!!

Our mac mail server mounts another mac servers raid share called BACKUPS, in this share there is a folder called KERIO BACKUP, where the kerio dumps it's zip files.

This in turn then get's backed up to the tape backup JIT.

Since i've updated to version 7.1.0(1906), i've had disk space issues where i've randomly run out of space on the server.

What is happening is that instead of it backing up the the mounted share, it is creating it's own folder of the same name at the root as the backup share, and then trying to dump the backup in there.

This is fine for almost a week, but then kills the local HD space, and needs deleting from /volumes, which incidently shows a network icon (globe) in the gui.

In a word WTF?

When a network volume is mounted, and not disconnected, like this one, X sees it in the volumes as a unique name! Why is Kerio ignoring this and then creating it's own folder in the root?

I've made sure that the full path is correct after the last time it did this in the daytime, and didn't disturb my Shiraz.

This must be fixed, else I fear my Shiraz will be spoiled!!
Re: I think i've found a backup issue/bug... [message #74975 is a reply to message #74973] Thu, 28 October 2010 03:10 Go to previous messageGo to next message
j.a.duke is currently offline  j.a.duke
Messages: 239
Registered: October 2006
lythaby wrote on Wed, 27 October 2010 18:07
I've just had to ssh in from home(Somewhat spoiling my bottle of Rosemount Shiraz), to delete the Kerio Backup Folder!!

Our mac mail server mounts another mac servers raid share called BACKUPS, in this share there is a folder called KERIO BACKUP, where the kerio dumps it's zip files.

This in turn then get's backed up to the tape backup JIT.

Since i've updated to version 7.1.0(1906), i've had disk space issues where i've randomly run out of space on the server.

What is happening is that instead of it backing up the the mounted share, it is creating it's own folder of the same name at the root as the backup share, and then trying to dump the backup in there.

This is fine for almost a week, but then kills the local HD space, and needs deleting from /volumes, which incidently shows a network icon (globe) in the gui.

In a word WTF?

When a network volume is mounted, and not disconnected, like this one, X sees it in the volumes as a unique name! Why is Kerio ignoring this and then creating it's own folder in the root?

I've made sure that the full path is correct after the last time it did this in the daytime, and didn't disturb my Shiraz.

This must be fixed, else I fear my Shiraz will be spoiled!!


Thanks for posting this-I'm having the same issue and just thought it was an issue with the remote AFP volume not mounting after the weekly restart (even though it's in the login items). Not to mention, I thought it was unique to my config.


I'd have to go back to pinpoint when it started, but I did update to 7.1.something on Labor Day, so that's about right for a timeframe.

I'm about to install a launchd agent to keep those volumes mounted to see if that resolves the problem.

I'll keep you posted. If you learn anything, let me know.

Thanks.

Cheers,
Jon
Re: I think i've found a backup issue/bug... [message #74978 is a reply to message #74973] Thu, 28 October 2010 10:37 Go to previous messageGo to next message
lythaby is currently offline  lythaby
Messages: 5
Registered: June 2007
Excellent, that makes perfect sense.

I've only rebooted twice since the upgrade, and have the network share in login items.

I'm assuming that Kerio is now wanting you to manually mount the share after reboot, and then re-find the path in the console.

Richard
Re: I think i've found a backup issue/bug... [message #74980 is a reply to message #74978] Thu, 28 October 2010 11:16 Go to previous messageGo to next message
Pavel Dobry (Kerio) is currently offline  Pavel Dobry (Kerio)
Messages: 2057
Registered: October 2003
Location: Czech Republic
lythaby wrote on Thu, 28 October 2010 10:37

I'm assuming that Kerio is now wanting you to manually mount the share after reboot, and then re-find the path in the console.
Richard


There was no change. If you're using a mounted network share for backup you must make sure that the disk is mounted before the Kerio Connect starts (so mounting in Login Items is too late). You need to use LaunchDaemon for this and make a script to define services dependency - mount first then start Kerio Connect. Alternatively you can add disk mounting to the Kerio Connect start-up script (and do it with every update).
Also the disk must remain mounted otherwise the backup will fail.

This requirement was there all the time.


Re: I think i've found a backup issue/bug... [message #74981 is a reply to message #74980] Thu, 28 October 2010 11:54 Go to previous messageGo to next message
lythaby is currently offline  lythaby
Messages: 5
Registered: June 2007
Although I don't want to appear argumentative, i really must make this point.

We've had KMS/Connect for about 4 years on this site, without issue, and only since upgrading to Connect have i started to see this problem.

Before then i was part of a Kerio reseller for about 3 years, so i have seen my fair share of installs/configs.

Hence maybe the issue isn't a point upgrade bug, but a version upgrade bug?

I'll have a play with the scripts.
Re: I think i've found a backup issue/bug... [message #74997 is a reply to message #74980] Thu, 28 October 2010 17:42 Go to previous messageGo to next message
j.a.duke is currently offline  j.a.duke
Messages: 239
Registered: October 2006
Kerio_pdobry wrote on Thu, 28 October 2010 05:16
lythaby wrote on Thu, 28 October 2010 10:37

I'm assuming that Kerio is now wanting you to manually mount the share after reboot, and then re-find the path in the console.
Richard


There was no change. If you're using a mounted network share for backup you must make sure that the disk is mounted before the Kerio Connect starts (so mounting in Login Items is too late). You need to use LaunchDaemon for this and make a script to define services dependency - mount first then start Kerio Connect. Alternatively you can add disk mounting to the Kerio Connect start-up script (and do it with every update).
Also the disk must remain mounted otherwise the backup will fail.

This requirement was there all the time.


I'd like to respectfully disagree with the real world experience I've seen.

Before 4 September (when I installed 7.1.0p1), mounting backup shares after login posed no problem with the backup being performed. This behavior has been consistent for at least the last 2 years.

After the update, the problems noted above have occurred. It would seem that Connect 7.1 now enforces the need for that volume to be present when the mailserver process starts, rather than the older behavior of the volume just needed to be mounted when the backup ran.

I don't recall any issues with the 6.x series of releases or with 7.0.x.

I don't have any problem with the volume needing to be present at the time of the backup, but the market at which Connect seems to be targeting might not have the wherewithal to craft the appropriate launchd items to ensure compliance with the more stringent requirements.

Cheers,
Jon
Re: I think i've found a backup issue/bug... [message #75074 is a reply to message #74997] Sun, 31 October 2010 02:55 Go to previous messageGo to next message
j.a.duke is currently offline  j.a.duke
Messages: 239
Registered: October 2006
I just reviewed my notes of what I did on 4 Sept and realized that I had upgraded the server to 10.6.4 at the same time from 10.4.11. I don't know if the OS version has any impact on Connect behavior regarding backup.

Richard, what OS are you running on your server? Intel or PowerPC (if <10.6)?

Also, I'd be happy to share my shell and launchd scripts with you.

Cheers,
Jon
Re: I think i've found a backup issue/bug... [message #75076 is a reply to message #75074] Sun, 31 October 2010 12:08 Go to previous messageGo to next message
Pavel Dobry (Kerio) is currently offline  Pavel Dobry (Kerio)
Messages: 2057
Registered: October 2003
Location: Czech Republic
j.a.duke wrote on Sun, 31 October 2010 02:55
I just reviewed my notes of what I did on 4 Sept and realized that I had upgraded the server to 10.6.4 at the same time from 10.4.11. I don't know if the OS version has any impact on Connect behavior regarding backup.

Cheers,
Jon


This is also the answer you're looking for. Mac OS X Tiger was using different startup scripts, which have dependencies (ie. Kerio Connect was started after all disks are mounted). Leopard and Snow Leopard use different system for starting programs and does not support using of dependencies.


Re: I think i've found a backup issue/bug... [message #75340 is a reply to message #75076] Fri, 05 November 2010 18:30 Go to previous messageGo to next message
Ron!
Messages: 1
Registered: April 2005
Location: Washington, DC
Pdobry: Was the startup scripts changed for leopard or snow leopard?

I am running 6.7.x (waiting for all these issue to be discovered and fixed) with leopard and backup daily to a AFP volume. Have not experienced this problem.

Re: I think i've found a backup issue/bug... [message #75447 is a reply to message #75340] Tue, 09 November 2010 16:24 Go to previous messageGo to next message
lythaby is currently offline  lythaby
Messages: 5
Registered: June 2007
Sorry for the delay.

We're on a PowerPC G5 Xserve, running 10.4.11.

I've removed the automated mounts, and will do it manually. Luckily, touch wood, i only need to re-boot them about 4-5 times a year.
Re: I think i've found a backup issue/bug... [message #80913 is a reply to message #74973] Tue, 28 June 2011 05:17 Go to previous messageGo to next message
mckingsley is currently offline  mckingsley
Messages: 37
Registered: April 2008
Hey I have run into this issue a couple times also, and I am not using a network mounted volume for the backup. I am using a plain ole usb external drive. If the drive is not there it should not backup, period. But I also had it taking up space on my boot volume.
Re: I think i've found a backup issue/bug... [message #103723 is a reply to message #80913] Wed, 12 June 2013 08:21 Go to previous messageGo to next message
Nielsvk is currently offline  Nielsvk
Messages: 5
Registered: June 2013
Running into the same problem is describe above.
I'm pretty sure that before 2012 a mounted AFP volume was never a problem.
It's started here somewhere in 2012 with no OSX updates.

What's the best simple fix so far? As i am not a script / Terminal guru.

[Updated on: Wed, 12 June 2013 08:23]

Report message to a moderator

Re: I think i've found a backup issue/bug... [message #103730 is a reply to message #74973] Wed, 12 June 2013 10:37 Go to previous message
lythaby is currently offline  lythaby
Messages: 5
Registered: June 2007
I just made a text file, words akin to "mount the backup server, or else", and added it to startup.

It now seems that as-long-as you have it mounted before the backup starts you can get away with it. Also remember that if you reboot the server that its mounted too, it will loose the connection.

It's no problem for me now, as I've moved to O365.
Previous Topic: Web based chat client
Next Topic: Public Folder only for 2 Users visible
Goto Forum:
  


Current Time: Mon May 29 17:53:30 CEST 2023

Total time taken to generate the page: 0.05553 seconds