Page 1 of 1

File Corruption using Ubuntu Server version 10.10

Posted: Mon Jan 31, 2011 8:32 am
by terrydlm
I have tried using Centos 32 and 64 bit but after coming up with the restrictions on the amount of events I can store (32,000 ish per camera) I found that I needed to use a Reiser file system.

I therefore installed Ubuntu Server 10.10 64 bit. This machine is ONLY used by Zoneminder.

I have installed Zonemidner version 1.24.2 and this generally works fine.

I have a 1 TB hard drive to store the events to and I have set the Purge When Full Filter to 80%.

I have 7 coax attached cameras using an 8 port BTTV card.

I have to admit that this is probably streching Zoneminder a bit (from what I have read people don't seem to run as many cameras at once) as we can fill up 800GB in just under 2 weeks however after some tweeking Zoneminder in the whole works fine.

The problem I have now since I am now using a Reiser fs is that when the events are being deleted (i think) some files become corrupted. The file attributes change to ? and the events cannot be deleted or modified. The system then becomes read only and a reboot is required so that it can be used again. As long as the corrupt file is not touched then the server is OK BUT there are jobs that run at midnight which obviosuly touch these files as in the morning the server is read only again.

The only way I have found to remove these items is to delete as much as I can from the events cache (leaving the corrupt files of course and rebooting every time I find one as the system is read only) and then boot up using a live cd and then run reiserfsck and rebuild-tree then boot back into zoneminder and off we go for another 2 weeks.

Does anyone have any ideas as to what is causing this corruption ? It did not happen with an ext3 file system but I can't use that because of the limit of 32000 ish events per camera which we soon reach.

Any help would be greatly appreciated as I am sure you can image deleting 800GB of data is not a fast process, especially when you have to keep rebooting when you find a corrupt file.

Posted: Mon Jan 31, 2011 9:02 am
by zoneminder
It doesn't directly address your problem but if you configure ZM to use deep file storage then it is unaffected by the 32k limit, plus runs much faster.

Unfortunately it probably won't help you too much as you can't really change from flat to deep and still keep your events but for a new install it is recommended (and will be the default from 1.24.3).

Posted: Mon Jan 31, 2011 9:17 am
by terrydlm
zoneminder wrote:It doesn't directly address your problem but if you configure ZM to use deep file storage then it is unaffected by the 32k limit, plus runs much faster.

Unfortunately it probably won't help you too much as you can't really change from flat to deep and still keep your events but for a new install it is recommended (and will be the default from 1.24.3).
Hi Zoneminder,

Thanks for your prompt response.

I would rather use Centos (Ext3 or 4) then Ubuntu (Reiser) so I think I will wait for Centos 6 to be released and Zoneminder 1.24.3 and then rebuild the server with the deep file storage on.

Thanks for your help.

Terry