[Info-vax] OpenVMS servers and clusters as a cloud service

[Simon Clubley]

On 2018-01-08, IanD <iloveopenvms at gmail.com> wrote:
> I think VMS clusters have been bypassed now
>

For some things, not for others.

> Instead of building robust clusters with shared architecture, the IT world
> built a multi threaded application layer instead
>
> Hadoop mitigates the need for robust clusters. Just fire up a stack of cheap
> Linux servers, deploy your workload across it, add however much redundancy you
> want to handle for failures among the way and your done
>

With Cassandra, I don't know how close you can get to the 100% data
integrity that VMS clusters give you, but I certainly don't like some
of the defaults:

https://stackoverflow.com/questions/31032156/cassandra-is-configured-to-lose-10-seconds-of-data-by-default

Oh and you had better not do a read with a ConsistencyLevel of less than
ALL unless you don't care about your data:

https://wiki.apache.org/cassandra/ReadRepair

It appears that HBase is supposed to be better at this than Cassandra
is, but it's still not fully clear to me how the HBase integrity compares
to VMS clusters when things go wrong.

>
> What good is it if I run a batch job for 36 hours on VMS if the node crashes
> and I have to start again. I've wasted 36 hours!
>

The standard solution for long running jobs on VMS and elsewhere is
to checkpoint the job at regular intervals.

Simon.

-- 
Simon Clubley, clubley at remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world