[ipv6calc] ipv6calc and Berkeley DB

Petr Kubat pkubat at redhat.com
Wed Jul 27 09:36:51 CEST 2016


Hello Peter,


On 07/25/2016 09:30 PM, Peter Bieringer wrote:
> Hi Petr,
>
> Am 25.07.2016 um 08:45 schrieb Petr Kubat:
>>> => what are the plans here for future Fedora and EL versions (e.g. EL 8)?
>> The current plan is to remove Berkeley DB dependencies by the time EL8
>> hits. We might be keeping the Berkeley DB packages in EL8 for the sake
>> of compatibility and for data migration tools however.
> If I read your status regarding LMDB, TokyoCabinet, KyotoCabinet, I
> believe EL8 can't drop Berkeley DB completly, because there is no
> alternative database is currently in EL6 or EL6 vanilla available.

As touched in the previous email, Berkeley DB will not be dropped 
completely in EL8 given that we need it to provide the users with a 
means of migration to new formats.

>
> And once EL8 would be released, EL6 is still under support, means
> software which should run on vanilla EL6+7+8 in compatible way has to
> use BerkeleyDB.
>
> Only chance would be to put LMDB in a later minor release to vanilla
> EL6+EL7 before EL8 will be released - are there plans for that?

This will not happen for EL6 as it is entering production phase 2. For 
EL7, it is possible that the package will be included if needed. Both 
KyotoCabinet and LMDB are already present in EPEL6 and EPEL7 repos.
Given that ipv6calc is located in Fedora and  EPEL repos only, you 
should not have any problems with using either LMDB or KyotoCabinet for 
the alternative to BDB. Apart from the Perl wrappers not being packaged 
at the moment but that should be easy enough to fix.

>
> And TokyoCabinet already reached somehow it's EOSL by the
> maintainer...so also not a good candidate for the future.
>
>
> Opinions about use of SQLite instead of BerkeleyDB?

Unfortunately I am not familiar with the details of SQLite. However, 
given that it is a different type of a data store (relational vs. 
key-data pair) I would say that the migration to SQLite would be a bit 
more difficult.

> Regards,
> 	Peter
>

Thanks,
Petr


More information about the ipv6calc mailing list