Anyone ever tried Data Extractor RAID Edition?

hddguy

New member
IT is defined at '0' but it is not recognized unless it is recognized under 'quick disk analysis' or if it is imported as virtual super-block when viewing in sector editor in DE.

What meta-data is used for auto-detection? How is user data used for auto detection?

Varied RAID controllers will have variety of meta-data, this data will also change when disks are inserted to new slots etc, or are you referring to LVM/LDM metadata? If RAID DE supports LVM/LDM then this could be useful but from experience I dont think it does?

The only real way to verify a RAID configuration without support from LVM type info is to identify and follow the pattern and length of XOR and calculate offset and order based on it. In some systems there is partition 'counters' that can be used - is this the detection by user data you mean?
 

acelab

New member
Hi everybody,

Thanks "jol". From time to time will be here and make some posts.

hddguy":26q7mca6 said:
IT is defined at '0' but it is not recognized unless it is recognized under 'quick disk analysis' or if it is imported as virtual super-block when viewing in sector editor in DE.
We’ve checked this case on the latest version. The file system has been successfuly recognized. Maybe there was an error in the version which you used. Anyway we can figure out it, if we are able to see this case.

hddguy":26q7mca6 said:
What meta-data is used for auto-detection?
In the latest version DE support these kinds of metadata: Windows LDM, mdadm (versions 0.9, 1.0, 1.1, 1.2), Promise FastTrack. If there’re several types of meta-data in one case the Data Extractor checks out all configurations. And we’re going to add support LVM, IntelMatrixRAID, DDF and some other kinds of meta-data. Some of them are identified and parsed in RAW Recovery now, so you can see information about RAID configuration.

hddguy":26q7mca6 said:
How is user data used for auto detection?
To cut in short there are two key points:
1) Using file system meta-data. For example, Data Extractor reads data from each member. Let's suppose that we’ve found MBR only on ‘Drive C’. Then we consider only configurations where Data block 0 is placed on ‘Drive C’. DE gets information about partitions from the MBR. Than DE searchs Boot/Superblock/VolumeHeader/.. for each partition. If it’s RAID0, then boot will be in one place. If it’s RAID5, then boot is located in another place. So when DE founds one boot, the number of possible configurations is reduced. If DE founds BootNTFS, it starts search for Boot copy, MFT0 and MFT Mirror. If DE founds XFS Superblok, it searchs for XFS Superblocks in another allocation groups. Sometimes the number of possible configurations is reduced.
2) Using user data. Data Extractor can check the correct size for many file types (the size of file part, that has no errors). When autodetection Is launched based on user data the Data Extractor tries to find block size and block order where the correct size will be the maximum. It allows to find the right configuration.

hddguy":26q7mca6 said:
Varied RAID controllers will have variety of meta-data, this data will also change when disks are inserted to new slots etc, or are you referring to LVM/LDM metadata? If RAID DE supports LVM/LDM then this could be useful but from experience I dont think it does?
It's difficult to say what happens when you replace slots in RAID array because of it depends on a type of RAID controller. But usually drives are identified in RAID, and slot replacement doesn't have to bear on RAID configuration.
We’ve listed all supported meta-data types above. All of them can be used at the same time. Of course there’re a lot of cases without any RAID metadata or with wrong meta-data. Using RAID meta-data is only one scenario to recognize RAID configuration, but it's not the only one possibility.

hddguy":26q7mca6 said:
The only real way to verify a RAID configuration without support from LVM type info is to identify and follow the pattern and length of XOR and calculate offset and order based on it. In some systems there is partition 'counters' that can be used - is this the detection by user data you mean?
Identifying XOR is just one possible example. What will you do if there’s no XOR? For example, configuration is RAID 0 or 1E with many drives. Or what will you do if XOR is a nonstandard? For example, RAID50 or RAID5 with delay.
If the case is difficult, we use the interactive mode (don’t confuse with autodetection modes, described above). It’s based on the same principles as autodetection by data. But it’s more powerful and flexible. And it’s more difficult to learn. The simple case is described in the article http://blog.acelaboratory.com/data-extr ... art-3.html, just see the video.
More difficult cases are shown on our Youtube channel http://www.youtube.com/playlist?list=PL ... o1X8E3DAvP . But keep in mind that some videos are almost three years old. There’re many improvements in the latests versions. Later on we will try to make new and fresh video tutorials about RAID.
 
Top