Failure of ndb_restore

[SHAWN HOGAN]

So this is really fun... We've been using ndbcluster fine for years and we had a catastrophic event at the data center that ultimately caused power failure to everything.

Long story short is we couldn't bring data nodes back online "as is" because of REDO log corruption.

We decided to just restore a backup from a few hours earlier, and now that is failing as well (not outright, but all sorts of error messages... Some of the errors ndb_restore are producing are:

[NdbApi] ERROR -- Sending TCROLLBACKREQ with Bad flag
---------------
theNoOfSentTransactions = 1 theListState = 0 theTransArrayIndex = 5
---------------
Temporary error: 410: REDO log files overloaded (decrease TimeBetweenLocalCheckpoints or increase NoOfFragmentLogFiles)
---------------
Temporary error: 266: Time-out in NDB, probably caused by deadlock
---------------


Ultimately making the restoration of the database impossible and ending nbd_restore at some point into it with:

-----
...Unable to recover from errors. Exiting...

NDBT_ProgramExit: 1 - Failed
-----

I've tried single user mode, I've tried --parallelism=1 (it seemed to recover "better" from some of the errors when *not* using it).

Using the latest GA version (7.3.5) on SuSE Linux Enterprise 11.

Any assistance would be great... at this point we just need to get back online.