[Rdo-list] PackStack Robustness across reboots

[Javier Pena]

----- Original Message ----- 

> Hi Haikel, I don't know if this counts as a bug or not but on Centos 7, after
> packstack is run successfully, "openstack-keystone" is not started [inactive
> (disabled on boot)]. The log for keystone shows the error "...Address
> already in use".
>
> Workaround ( source ) :
> Apparently httpd reserves the ports 5000 & 35357 during the packstack setup.
> So the ports (Listen 5000 & Listen 35357) have to be commented out in
> "/etc/http/conf/ports.conf"
> Then both httpd & keystone service have to be restarted using systemctl.
>
> Result:
> openstack-status shows openstack-keystone to be active

Actually, this is not a Packstack bug. 

Packstack configures Keystone to run inside httpd as a WSGI service, but openstack-status does not notice and keeps saying it is not started. The workaround steps you mentioned only revert that and make Keystone run as a standalone eventlet service.

Regards,
Javier

> On 5 August 2015 at 19:49, Haïkel < hguemar at fedoraproject.org > wrote:
>
> > 2015-08-05 19:29 GMT+02:00 sad man < asadxflow at gmail.com >:
> 
> > > Hi, some are simple other complex e.g., rabbit-mq server fail can be
> > > solved
> 
> > > simply by running systemctl start rabbitmq-server. Nova, neutron services
> 
> > > are a bit harder (I am doing some testing and short listing services that
> 
> > > fail often). Currently most common failures are :
> 
> > >
> 
> > > neutron-server
> 
> > > ceilometer-api
> 
> > > rabbitmq-server
> 
> > > (sometimes) horizon
> 
> > >
> 
> > > Let me fetch links for fixes, as I mentioned these failures are not
> 
> > > consistent.
> 
> > >
> 

> > AFAIK, this could be solved by fixing their respective systemd unit files.
> 
> > If you have patches, let me apply them.
> 

> > H.
> 

> > > On 5 August 2015 at 19:05, Haïkel < hguemar at fedoraproject.org > wrote:
> 
> > >>
> 
> > >> 2015-08-05 18:02 GMT+02:00 sad man < asadxflow at gmail.com >:
> 
> > >> > Hi RDO folks, I need some suggestions regarding packstack + rdo
> 
> > >> > robustness.
> 
> > >> > I have integrated the option to install RDO into CentOS installer and
> 
> > >> > "openstack-status" shows all services to be either active or inactive
> 
> > >> > (which
> 
> > >> > I think means not "failed").
> 
> > >> >
> 
> > >> > But on reboot some services randomly fail to restart e.g., horizon,
> 
> > >> > nova-network etc. are shown as "failed" after some reboots. There are
> 
> > >> > some
> 
> > >> > "fixes" on the forums to restart these services.
> 
> > >> >
> 
> > >> > I wanted some suggestions on how I can make a "robust" openstack setup
> 
> > >> > (I am
> 
> > >> > using simplest "packstack --allinone") which doesn't fail.
> 
> > >> >
> 
> > >> >
> 
> > >> > PS: Right now I am thinking of scripting these "fixes" and checking
> > >> > the
> 
> > >> > openstack-status on each reboot through systemd.
> 
> > >> >
> 
> > >>
> 
> > >> Could you point me these fixes?
> 
> > >>
> 
> > >>
> 
> > >> > --
> 
> > >> > Cheers,
> 
> > >> >
> 
> > >> > Asad
> 
> > >> >
> 
> > >> > _______________________________________________
> 
> > >> > Rdo-list mailing list
> 
> > >> > Rdo-list at redhat.com
> 
> > >> > https://www.redhat.com/mailman/listinfo/rdo-list
> 
> > >> >
> 
> > >> > To unsubscribe: rdo-list-unsubscribe at redhat.com
> 
> > >
> 
> > >
> 
> > >
> 
> > >
> 
> > > --
> 
> > > Cheers,
> 
> > >
> 
> > > Asadullah Hussain
> 

> --
> Cheers,

> Asadullah Hussain

> _______________________________________________
> Rdo-list mailing list
> Rdo-list at redhat.com
> https://www.redhat.com/mailman/listinfo/rdo-list

> To unsubscribe: rdo-list-unsubscribe at redhat.com