OpenVZ Forum


Home » General » Support » large ploop files (worried about corruption)
large ploop files (worried about corruption) [message #51533] Thu, 10 July 2014 14:58 Go to next message
vmvmvm is currently offline  vmvmvm
Messages: 51
Registered: January 2006
Member
Hello,

I am planning to convert some simfs to ploop.

The simfs are from 100GB to 500GB.

Will these work as ploop?

Do I need to worry about corruption? I'm afraid that any physical disk issue might cause the whole ploop file image to be unreadable. With simfs any disk issue would be local to a few files in the best case.

Thanks for your advice!

Re: large ploop files (worried about corruption) [message #51535 is a reply to message #51533] Thu, 10 July 2014 19:17 Go to previous messageGo to next message
devonblzx is currently offline  devonblzx
Messages: 127
Registered: December 2006
Senior Member
Nearly every virtualization technology accepts images as loop devices.

We have 1TB+ servers on ploop images if that makes you feel better. You can run fsck on the ploop device in case of filesystem issues.

If you are talking about a disk failure then simfs isn't going to save you from much. Most times a disk fails, you'd be lucky to get a few files off of it, if any at all. Using RAID mirroring on the host node and monitoring the RAID status will help avoid that.


http://static.openvz.org/userbars/openvz-user-2.png
ByteOnSite President
Re: large ploop files (worried about corruption) [message #51536 is a reply to message #51535] Thu, 10 July 2014 19:26 Go to previous messageGo to next message
vmvmvm is currently offline  vmvmvm
Messages: 51
Registered: January 2006
Member
Hi,

Thanks for your reply.

Regarding the ploop images - I may have a technical mis-understanding. My fear is that if a few bytes of the image are corrupted, the entire file system would become bad?

I am no doubt moving to ploop - there are many advantages. Just a bit paranoid about moving really large filesystems to it.

Thanks again!
Re: large ploop files (worried about corruption) [message #51720 is a reply to message #51533] Wed, 22 October 2014 12:41 Go to previous message
ccto is currently offline  ccto
Messages: 61
Registered: October 2005
Member
Hello all,

May I ask your experience that if a hardware node, equipped with mdadm RAID-1, hang/crashed (for whatever reason, e.g. power failure), then when it reboots, it will fsck.ext4 the host file system. Do you encounter issue for the guest ploop file system?

I saw some post in webhostingforum.com that it may corrupt the ploop file system
http://www.webhostingtalk.com/showthread.php?t=1282075&p age=3

or simply fsck.ext4 on ploop file system, then it is OK again.

Thank you very much for your attention.

Regards
George
Previous Topic: Lame question on fresh install of openvz
Next Topic: ip_conv: command not found (Debian)
Goto Forum:
  


Current Time: Sun Nov 17 22:41:26 GMT 2024

Total time taken to generate the page: 0.02884 seconds