Re: [Jdbm-general] Question

[Alex B. <boi...@in...>]

Eric,

This looks like a file corruption problem.  Each block allocated by JDBM 
is marked with a "magic" two-byte marker to allow reconstruction and 
introspection of the database.

In this case, the magic number is invalid (value 0) instead of being 
0x1353 (Magic.BLOCK + Magic.TRANSLATION_PAGE).

I can't really speculate on the source of corruption without further 
information.

alex

Rosenberg, Eric wrote:
> I just got the following error which I have never seen before. What types of
> things could have caused the jdbm files to get into this state?
> 
> java.lang.Error: CRITICAL: page header magic for block 5 not OK 0
> 	at jdbm.recman.PageHeader.<init>(PageHeader.java:75)
> 	at jdbm.recman.PageHeader.getView(PageHeader.java:98)
> 	at jdbm.recman.PageManager.getNext(PageManager.java:190)
> 	at jdbm.recman.PageCursor.next(PageCursor.java:92)
> 	at
> jdbm.recman.PhysicalRowIdManager.fetch(PhysicalRowIdManager.java:158)
> 	at jdbm.recman.BaseRecordManager.fetch(BaseRecordManager.java:325)
> 	at jdbm.recman.CacheRecordManager.fetch(CacheRecordManager.java:251)
> 	at jdbm.recman.CacheRecordManager.fetch(CacheRecordManager.java:230)
> 
> Eric Rosenberg