Re: DB corruption (double get for block) fixed in M25 or upcoming M26?

2019-09-11 Thread Emmanuel Lécharny

Hi Mike,


I checked, and what has been added is the support of the Transaction 
extended operation in the server. It obviously uses the internal 
transaction system in plave in M25, so I don't think it will fix the 
issue you are facing.


Bottom line, we really have to get rid if JDBM - the internal DB 
library-  which seems to be unreliable. Some effort has been started 
years ago, but stalled, with Mavibot. We are no that far from having a 
solid new DB, but I can't promise to be able to have it running before 
the end of this year...



On 11/09/2019 21:32, Michael Grove wrote:

Thanks for the response.

  


Re:

  


Now, and I have to check, I *think* we have added transaction support at

a higher level, and it may be only present in M26 (which is to be

released soon).

  


Did you get a chance to determine if transaction support has been added at a
higher level in M26?  If it hasn't, am I correct that there's not an
expectation that M26 fixes this issue?

  


Thanks.

  


-mike




-
To unsubscribe, e-mail: users-unsubscr...@directory.apache.org
For additional commands, e-mail: users-h...@directory.apache.org



RE: DB corruption (double get for block) fixed in M25 or upcoming M26?

2019-09-11 Thread Michael Grove
Thanks for the response.

 

Re:

 

Now, and I have to check, I *think* we have added transaction support at 

a higher level, and it may be only present in M26 (which is to be 

released soon).

 

Did you get a chance to determine if transaction support has been added at a
higher level in M26?  If it hasn't, am I correct that there's not an
expectation that M26 fixes this issue?

 

Thanks.

 

-mike