[Info-vax] stopping and starting failover queues

[Phillip Helbig undress to reply]

In article <4fae5027-c1ad-4a1c-b043-764f26884f4fn at googlegroups.com>,
Edgar Ulloa <ulloa.edgar at gmail.com> writes: 

> El lunes, 16 de agosto de 2021 a las 5:54:08 UTC-5, Phillip Helbig (undress=
>  to reply) escribi=C3=B3:
> > Consider a queue which can run on several nodes for failover. However,=20
> > if all nodes are up, one node might be preferred. How can one force the=
> =20
> > jobs in it to run on the preferred node? One can do stop/ queu /on on=20
> > the node where it is running, get it to fail over (repeat if necessary),=
> =20
> > then do ENABLE AUTOSTART/QUEUES and/or START/QUEUE where it was stopped=
> =20
> > (to get all queues running again). However, there must be a better way.
> 
> Maybe it's not the best way but it works for me like this
> 
> Instead of sending it to the generic queue that contains the various nodes=
> =20
> I made the queue on the dedicated server and with that only the jobs sent t=
> o that queue will be executed on the dedicated node.
> regards

There is no generic queue pointing to more than one specific queue; it 
is a single queue on more than one node.  The former is good for 
failover (if one SUBMITs and programs correctly) and load distribution; 
the latter is only for failover and easier to implement.