Create the ppc.fedoraproject.org DNS zone Create the 10.5.129.0 reverse lookup zone
Add the following initial records:
{{{ ppc-hub01.ppc.fedoraproject.org 10.5.129.1 arm-db01.qa.fedoraproject.org 10.5.129.2 compose-ppc64-01.ppc.fedoraproject.org 10.5.129.4 compose-ppc64le-01.ppc.fedoraproject.org 10.5.129.5
ppc8-01.qa.fedoraproject.org 10.5.129.10 ppc8-02.qa.fedoraproject.org 10.5.129.11 ppc8-03.qa.fedoraproject.org 10.5.129.11 ppc8-04.qa.fedoraproject.org 10.5.129.12 }}}
I modified this request some (no qa in names, we reserve the first 10 ips, got rid of duplicate ips, and named the hub/db the same way we name all the other ones), which gives:
{{{ ppc-koji01.ppc.fedoraproject.org 10.5.129.11 db-ppc-koji01.ppc.fedoraproject.org 10.5.129.12 compose-ppc64-01.ppc.fedoraproject.org 10.5.129.14 compose-ppc64le-01.ppc.fedoraproject.org 10.5.129.15
ppc8-01.ppc.fedoraproject.org 10.5.129.20 ppc8-02.ppc.fedoraproject.org 10.5.129.21 ppc8-03.ppc.fedoraproject.org 10.5.129.22 ppc8-04.ppc.fedoraproject.org 10.5.129.23 }}}
If you need adjustments in those or have more, just reopen and add em and we can get them added.
Can I also have:
{{{ buildvm-ppc64-01.ppc.fedoraproject.org 10.5.129.64 buildvm-ppc64le-01.ppc.fedoraproject.org 10.5.129.65 }}}
Can we also have RW access for the following hosts to fedora_ppc for the following on either nfs specific addresses or the above addresses. If the later please assign IPs allocate DNS: {{{ ppc-koji01.ppc.fedoraproject.org compose-ppc64-01.ppc.fedoraproject.org compose-ppc64le-01.ppc.fedoraproject.org buildvm-ppc64-01.ppc.fedoraproject.org buildvm-ppc64le-01.ppc.fedoraproject.org }}}
Can we also generate fedmsg certs for the hub and compose boxes (not sure if the runroot koji builder hosts need them too, if so the last two too).
Can we also have IPs allocated for OOBM and assoicated DNS added (not sure if mgmt is on a separate sub domain). mgmt hostnames are:
{{{ ppc8-01-fsp ppc8-02-fsp ppc8-03-fsp ppc8-04-fsp }}}
Added the new buildvms to dns.
The fsps are:
10.5.130.50, 51, 52, 53
I setup nfs exports just using the 10.5.129 ips for now.
Lets check with Ralph on monday about the certs, I am not sure which are needed for which host off hand. Will leave this open for that.
Update the configs for the ASMs. MAC addresses for each of the spare ASM interfaces are as follows:
{{{ ppc8-01-fsp 10.5.130.50 40:F2:E9:A5:47:60 ppc8-02-fsp 10.5.130.51 98:BE:94:4B:96:10 ppc8-03-fsp 10.5.130.52 98:BE:94:4B:94:C4 ppc8-04-fsp 10.5.130.53 98:BE:94:4B:94:54 }}}
Ah, I already added those in dns, so unless they need dhcp that should be all set. ;)
Whats left here? Did the certs get sorted out?
Replying to [comment:6 kevin]:
Ah, I already added those in dns, so unless they need dhcp that should be all set. ;) Whats left here? Did the certs get sorted out?
I'm not sure the new FSP NICs have been patched yet, I can't ping any of the above IPs. I'll be rebuilding the hosts once we come out of freeze and have ppc Alpha out so I'll have some new builder DNS to do then to.
They were connected, but the switch ports haven't been set yet.
I just put in a ticket on that and will update here when it's complete.
ok. The ports should now be set.
Please see if you can enable them now.
Can I have the following new ppc.fedoraproject.org DNS records:
{{{ buildvm-ppc64-02.ppc.fedoraproject.org 10.5.129.66 buildvm-ppc64-03.ppc.fedoraproject.org 10.5.129.67 buildvm-ppc64-04.ppc.fedoraproject.org 10.5.129.68 buildvm-ppc64-05.ppc.fedoraproject.org 10.5.129.69 buildvm-ppc64-06.ppc.fedoraproject.org 10.5.129.70 buildvm-ppc64-07.ppc.fedoraproject.org 10.5.129.71 buildvm-ppc64-08.ppc.fedoraproject.org 10.5.129.72 buildvm-ppc64le-02.ppc.fedoraproject.org 10.5.129.73 buildvm-ppc64le-03.ppc.fedoraproject.org 10.5.129.74 buildvm-ppc64le-04.ppc.fedoraproject.org 10.5.129.75 buildvm-ppc64le-05.ppc.fedoraproject.org 10.5.129.76 buildvm-ppc64le-06.ppc.fedoraproject.org 10.5.129.77 buildvm-ppc64le-07.ppc.fedoraproject.org 10.5.129.78 buildvm-ppc64le-08.ppc.fedoraproject.org 10.5.129.79 }}}
done.
Also for ppc.fedoraproject.org:
{{{ ppc-koji01 10.5.129.240 db-ppc-koji01 10.5.129.241 }}}
Replying to [comment:10 kevin]:
ok. The ports should now be set. Please see if you can enable them now.
So it looks like they've broken the new secondary VLAN that is patched into these hosts:
I can't ping these hosts, and rebooting them makes no difference and they're all accessible on the virtual consoles. The two koji hosts stopped checking in 2016-03-31 21:32:59, we need this fixed ASAP because it's broken all PPC composes.
buildvm-ppc64-01.ppc.fedoraproject.org buildvm-ppc64le-01.ppc.fedoraproject.org compose-ppc64-01.ppc.fedoraproject.org compose-ppc64le-01.ppc.fedoraproject.org
Put in a update to the ticket to revert the switch changes that caused this. :(
Can I have RO NFS for the following hosts (for createrepo). We'll soon be able to clean up the old ones.
Replying to [comment:17 pbrobinson]:
RO only NFS for fedora_ppc/data (if that wasn't already clear)
Done, they have access now
So outstanding on this ticket:
There's a back view of the Power 8 chassis on page 19 section "1.4 Physical package", Figure 1.3 http://www.redbooks.ibm.com/redpapers/pdfs/redp5098.pdf
There are two RJ-45 connectors bottom left, one should be already patched. The other one needs to be. It would be useful to know the server port/switch port that the exiting ones is patched into (we need remove them off the QA vlan later)
Reconfigure old QA interfaces to new secondary VLAN interface (will give switch ports and add here)
{{{ VLAN: 684
ppc8-01: eth0: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/20 ppc8-02: eth4: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/26 ppc8-03: eth0: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/17 ppc8-04: eth0: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/23 }}}
Replying to [comment:21 pbrobinson]:
Reconfigure old QA interfaces to new secondary VLAN interface (will give switch ports and add here) {{{ VLAN: 684 ppc8-01: eth0: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/20 ppc8-02: eth4: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/26 ppc8-03: eth0: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/17 ppc8-04: eth0: sw01-access-rack17.mgmt.phx2.redhat.com: GigabitEthernet0/23 }}}
Do we even need to change these? Or were you planning on adding them to the bridge to have 2 interfaces instead of the existing one?
I put in a ticket on the mgmt ports again, hopefully we will get a good answer this time, if not it may need to wait for our Q2 visit.
I was considering it, would likely be useful to balance the traffic in some way,but primarily I at least want them off the QA network so we don't get in QAs way.
OK, thanks
Internal ticket filed for the port changes.
Internal ticket for port changes done.
The iLO interface connection ticket is not yet done.
The hub/db move I am going to try and do tomorrow morning.
The iLO interfaces should be now done. Can you confirm they are live and on the right vlan?
(I think thats the last thing here, so we can close after thats confirmed?)
Replying to [comment:26 kevin]:
Sorry, missed this last update. Confirmed I can login via the new fsp/iLO IPs from the VPN so all looks good there (could only ping .50 from batcave so will poke that and try and work out when I get a moment) but I think we can close this off.
Login to comment on this ticket.