[Info-vax] MariaDB (Re: OT: what is old is new again?)
David Froble
davef at tsoft-inc.com
Wed Nov 23 15:05:08 EST 2016
Neil Rieck wrote:
> On Wednesday, May 4, 2016 at 4:52:17 PM UTC-4, David Froble wrote:
>> Craig A. Berry wrote:
>>> On 5/1/16 1:48 PM, Neil Rieck wrote:
>>>
>>>> We're currently running MariaDB-5.5-25 which can outperform ISAM RMS
>>>> They give it away but sell support
>>>> contracts. But they do not offer something for OpenVMS which means we
>>>> must rely on someone "to provide us with a build" or "do a build
>>>> ourselves". Now I asked that single provider if he was willing to sell
>>>> us a support contract and responded as not being interested in doing
>>>> this.
>>>
>>>> So for me I can only see two options.
>>>>
>>>> 1) develop a fallback strategy based upon Mimer
>>> >
>>>> 2) hope that someone at VSI is considering doing a MariaDB build
>>>
>>> And of course "doing a build" would be of limited help with your
>>> situation unless they were also going to offer support for it, right?
>>>
>>> At a TUD a few years ago, someone asked about OpenVMS Engineering
>>> providing support for MySQL and the HP person present said, "Why do you
>>> want us to do it?". Apparently that was a rhetorical question as he
>>> didn't wait for an answer. It seemed pretty obvious to me that the
>>> reason people want support for open source packages from the purveyor of
>>> VMS is that no one else is offering it.
>> Well, it seems like an opportunity to me. For example, I believe it's Mark
>> Berryman who is building MarieDB for VMS. If VSI decided to distribute the
>> product as part of VMS, and partnered with Mark to provide support, and made
>> sure customers had easy methods of obtaining such support, ...
>>
>> The "vendor" would perhaps have a sufficient revenue stream from support to
>> insure the resources for such support.
>>
>> Or, SQlite, or Postgre, or ...
>>
>> The key would be for VSI to pick one or more RDBMS products and make it happen.
>
> Let me second those remarks:
>
> First off, while I have been using Mark's port of MariaDB-5.5-25 since 2014,
we have noticed a few issues. When you check the blogs most people say something
like "Oh, that problem went away when we upgraded to MariaDB-10.1"
>
> I understand Mark works for a division of HP/HPE that is being outsourced so
he is very busy doing work that results in putting food on the table. That said,
with no version of MariaDB-10 on the horizon for OpenVMS, I have resorted to
installing CentOS Linux on a nearby server, then installing MariaDB-10.1.19
which we access from our OpenVMS development platform via the network. (think of
this experiment as an alternative form of SAN; I did this only to see if our
issues just went away, which they did)
>
> I'm not certain what the open source community did between MariaDB-5 and
MariaDB-10 but everyone claims it is faster and we noticed this as well.
Isn't that part of the reason for new versions, to fix problems?
> We are at a point where 50% of our OpenVMS data still remains on RMS and we
would like to move it to MariaDB but do not want to be trapped in a
single-source situation. (which VSI could solve by including MariaDB with
OpenVMS;
Now, this is where people like me, third party vendors, get confused. Whether
it's Mark, or someone else, or VSI, what's the difference? If a third party
agreed to provide support and such, they why do you insist on only the developer
of VMS?
> or we could solve by sticking with the Linux alternative); our only
other alternative is to investigate something like Mimer:
> http://developer.mimer.com/platforms/index.tml ) >
> Neil Rieck
> Waterloo, Ontario, Canada.
>
> p.s. do not think that SQLite is a viable alternative to MySQL, MariaDB or Postgres. While SQLite it is very powerful from a single user perspective, it introduces some new issues when used in a multi-user environment. That said, I play with it every day and believe it to be a viable alternative to small RMS-based ISAM projects. I have posted my efforts here:
> http://www3.sympatico.ca/n.rieck/docs/openvms_notes_sqlite.html
More information about the Info-vax
mailing list