On Wed, Dec 6, 2017 at 1:23 PM, Roberto Nunin <robnunin@gmail.com> wrote:
I've read the Bugzilla report, but after modify answer file, how to provide it to the Cockpit process ? I know that in the CLI installation it could be provided as aparameter of the CLI command, but in Cockpit.
You are right, unfortunately we have to fix it: I don't see any possible workaround at cockpit level. Directly running hosted-engine-setup from CLI instead should work.
2017-12-06 12:25 GMT+01:00 Yedidyah Bar David <didi@redhat.com>:
On Wed, Dec 6, 2017 at 1:10 PM, Roberto Nunin <robnunin@gmail.com> wrote:
This is the first request, note that do not allow to add CIDR: [image: Immagine incorporata 1]
2017-12-06 11:54 GMT+01:00 Roberto Nunin <robnunin@gmail.com>:
Yes, both times on Cockpit.
2017-12-06 11:43 GMT+01:00 Simone Tiraboschi <stirabos@redhat.com>:
On Wed, Dec 6, 2017 at 11:38 AM, Roberto Nunin <robnunin@gmail.com> wrote:
Ciao Simone thanks for really quick answer.
2017-12-06 11:05 GMT+01:00 Simone Tiraboschi <stirabos@redhat.com>:
> Ciao Roberto, > > On Wed, Dec 6, 2017 at 10:02 AM, Roberto Nunin <robnunin@gmail.com> > wrote: > >> I'm having trouble to deploy one three host hyperconverged lab >> using iso-image named above. >> > > Please note that ovirt-node-ng-installer-ovirt-4.2-pre-2017120512 > <%28201%29%20712-0512>.iso is still a pre-release software. > Your contribute testing it is really appreciated! >
It's a pleasure !.
> > > >> >> My test environment is based on HPE BL680cG7 blade servers. >> These servers has 6 physical 10GB network interfaces (flexNIC), >> each one with four profiles (ethernet,FCoE,iSCSI,etc). >> >> I choose one of these six phys interfaces (enp5s0f0) and assigned >> it a static IPv4 address, for each node. >> >> After node reboot, interface ONBOOT param is still set to no. >> Changed via iLO interface to yes and restarted network. Fine. >> >> After gluster setup, with gdeploy script under Cockpit interface, >> avoiding errors coming from : >> /usr/share/gdepply/scripts/blacklist_all_disks.sh, start >> hosted-engine deploy. >> >> With the new version, I'm having an error never seen before: >> >> The Engine VM (10.114.60.117) and this host (10.114.60.134/24) >> will not be in the same IP subnet. Static routing configuration are not >> supported on automatic VM configuration. >> Failed to execute stage 'Environment customization': The Engine VM >> (10.114.60.117) and this host (10.114.60.134/24) will not be in >> the same IP subnet. Static routing configuration are not supported on >> automatic VM configuration. >> Hosted Engine deployment failed. >> >> There's no input field for HE subnet mask. Anyway in our class-c >> ovirt management network these ARE in the same subnet. >> How to recover from this ? I cannot add /24 CIDR in HE Static IP >> address field, it isn't allowed. >> >
Incidentally, we also got now another similar report:
https://bugzilla.redhat.com/show_bug.cgi?id=1522712
Perhaps a new regression? Although I can't see where it happened.
> 10.114.60.117 and 10.114.60.134/24 are in the same IPv4 /24 subnet > so it should't fail. > The issue here seams different: > > From hosted-engine-setup log I see that you passed the VM IP address > via answerfile: > 2017-12-06 09:14:30,195+0100 DEBUG otopi.context > context.dumpEnvironment:831 ENV OVEHOSTED_VM/cloudinitVMStatic > CIDR=str:'10.114.60.117' > > while the right syntax should be: > OVEHOSTED_VM/cloudinitVMStaticCIDR=str:10.114.60.117/24 > > Did you wrote the answerfile by yourself or did you entered the IP > address in the cockpit wizard? if so we probably have a regression there. >
I've inserted it while providing data for setup, using Cockpit interface. Tried to add CIDR (/24), but it isn't allowed from Cockpit web interface. No manual update of answer file.
> > > >> >> Moreover, VM FQDN is asked two times during the deploy process. >> It's correct ? >> > > No, I don't think so but I don't see it from you logs. > Could you please explain it? >
Yes: first time is requested during initial setup of HE VM deploy
The second one, instead, is asked (at least to me ) in this step, after initial setup:
So both on cockpit side?
[image: Immagine incorporata 1]
> > >> >> Some additional, general questions: >> NetworkManager: must be disabled deploying HCI solution ? In my >> attempt, wasn't disabled. >> > Simone, could you confirm or not that NM must stay in place while deploying ? This qustion was struggling since 3.6...... what is the "best practice" ? All of mine RHV environments (3.6 - 4.0.1 - 4.1.6 ) have it disabled, but I wasn't able to find any mandatory rule.
In early 3.6 you had to disable it but now you can safely keep it on.
> There's some document to follow to perform a correct deploy ? >> Is this one still "valid" ? : https://ovirt.org/blog/2017/ >> 04/up-and-running-with-ovirt-4.1-and-gluster-storage/ >> >> Attached hosted-engine-setup log. >> TIA >> >> >> >> -- >> Roberto >> 110-006-970 >> >> _______________________________________________ >> Users mailing list >> Users@ovirt.org >> http://lists.ovirt.org/mailman/listinfo/users >> >> >
-- Roberto Nunin
-- Roberto Nunin
-- Roberto Nunin
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
-- Didi
-- Roberto Nunin
