Welcome, Guest. Please login or register.
Did you miss your activation email?
May 21, 2013, 12:13:26 PM
Home Help Search Login Register
News: Before posting, please read sticky topics!

+  DIY DataRecovery.nl Support forum
|-+  Support
| |-+  iRecover (Moderators: Tom, Joep)
| | |-+  RAID5 restore fails at 89% doing numerical analysis with out of memory error
0 Members and 1 Guest are viewing this topic. « previous next »
Pages: [1] Go Down Print
Author Topic: RAID5 restore fails at 89% doing numerical analysis with out of memory error  (Read 832 times)
Doug
member

Posts: 1


« on: December 28, 2009, 05:45:35 PM »

I am restoring a 1.6TB raid5 array on a windows 2003 server. irecover fails with a thread out of memory error toward the very end of the process (before displaying the files and folders). There are 6.5 to 16.5 million files on the array to be recovered. There are a few unrecoverable spots on the array, but most looks intact. There is one bad area on the NTFS directory so I expect some lost data.

If I image the drives to another server (a Windows 2008 web server) with 4 gig or more of memory, will the out of memory error go away, or is the out of memory error based on the size of some element within the iRecovery software? Is there a better way to run this restore that will not have the memory error? (I had already decreased used memory to about 200 meg before starting the process on the Windows 2003 server when the process failed. I saw the memory use pass 2.5 gb in hte evening and it failed overnight somewhere around 3 gb of use.

It's a few day process to image the data and at least a 24 hour period to regenerate the iRestore process on the new server.

Thanks in advance for your help,
Doug
856-478-2277
Logged
Tom
Developer and Support Tech
Administrator
member
*****
Posts: 1145


WWW
« Reply #1 on: December 28, 2009, 11:36:37 PM »

Quote
If I image the drives to another server (a Windows 2008 web server) with 4 gig or more of memory, will the out of memory error go away
Quite possible, yes. But not guaranteed.
Quote
or is the out of memory error based on the size of some element within the iRecovery software?
This i would have to look up, but for now i'd say no. In general, the limiting factor is actual physical memory.
Quote
Is there a better way to run this restore that will not have the memory error?
There are a few things you can do to save some memory, but with this many files i'm not sure it will make a difference. Go to "advanced options" and:
- enforce the file system type (set it to NTFS, assuming that's the file system type)
- disable the quick scan
- disable NTFS 1.2 compatibility
- and, obviously, disable any task or process that isn't needed while running iRecover

How's the array doing? Are you scanning x individual disks and having iRecover rebuild the array, or is the array itself intact and are you scanning the volume?
Logged

Joep
Developer and Support Tech
Administrator
member
*****
Posts: 1155


WWW
« Reply #2 on: December 29, 2009, 10:47:29 AM »

Hello,

The limiting factor is the amount of entries (files + directories) we need to process. Adding physical memory is recommended to avoid swapping but it will not make the error go away.

As Tom suggested close processes/tasks that aren't required but that tend to run in the background (virus scanners and all that). Slide down cache sizes in the runtime control as much as possible. Disable NTFS 1.2 compatibilty. Disable autosave.

To make more memory available to an individual process (more than 2 GB) add the /3GB switch in the boot.ini (http://technet.microsoft.com/en-us/library/bb124810(EXCHG.65).aspx).

We have tested upto 11.000.000 files/folders, peak memory usage was 2600 MB at that time. As a 32 bit process we will never be able to 'use' more than 3000 MB so with too many files/folders we will just run out of memory.

Logged

--
Kind regards,
Joep
Pages: [1] Go Up Print 
« previous next »
Jump to:  


Login with username, password and session length

Powered by MySQL Powered by PHP Powered by SMF 1.1.17 | SMF © 2011, Simple Machines Valid XHTML 1.0! Valid CSS!