[erlang-questions] Mnesia tables go out of sync in one node?

Tamas Nagy <>
Mon Sep 15 12:35:17 CEST 2008


Hi Everyone!

If this is really an mnesia bug this a major flow, because this breaks  
atomicity.
Could anyone advise on how to progress where to look?

Regards,
	Tamas

On 14 Sep 2008, at 19:46, Tamas Nagy wrote:

> Hi Mnesia Gurus!
>
> In one of my projects running mnesia:restore/1 seemingly finishes
> perfectly.
> The data is usable after restoration. But if I restart the node some
> tables loose a lot of data without any warning.
>
> Because the problem is reproducible I did some further testing and it
> seems that the data which disappears is never written to the disc
> that's why it disappears after a node restart.
>
> Is there any know limitation how big file mnesia:restore/1 can handle?
> Mine is around 300 Mb.
> Did the memory and disc data got of sync for anybody else before (not
> just during restoration)?
>
> The system is using R12B1 and consist of only one node. I've checked
> the release notes for R12B2 - R12B4 and there is nothing about this.
>
> As a side note: Does anyone now which hash function is used for
> building index tables? We had to change our unique identifier because
> building the index table for 1000000 rows took 2 and a half hours.
>
> Regards,
> 	Tamas
> _______________________________________________
> erlang-questions mailing list
> 
> http://www.erlang.org/mailman/listinfo/erlang-questions




More information about the erlang-questions mailing list