Closed Bug 1437144 Opened 7 years ago Closed 7 years ago

[MDC2] Install and configure deploystudio on install.test.releng.mdc2

Categories

(Infrastructure & Operations :: RelOps: General, task)

task
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: dividehex, Assigned: dividehex)

References

Details

Once the mac mini is racked and ported on the MDC2 virtual network, I'll need to manually install, configure and test the latest version of Deploystudio including the Apple Server app. The mini is currently on the SCL3 test network where it has been installed with Yosemite and puppetized. :van noted via IRC, we are waiting on Netops to configure the port before moving it to MDC2 virtual.
Assignee: relops → jwatkins
Status: NEW → ASSIGNED
For the most part, the mdc2 deploystudio server is installed and configured. I attempted to netboot one of the test minis that was moved to mdc2 (t-yosemite-r7-235.test.releng.mdc2.mozilla.com) but it got hung up somewhere in the process. I'll need :van to take a look at what is going on with it. Below are the general steps it took to set up the DS server: * Infoblox host record added (install.test.releng.mdc2.mozilla.com) * HostName, ComputerName, Host values changed via scutil * Deploystudio 1.7.8 installed and configured * Apple Server App (version 4.1.3) copied from MDC1 install (/Applications/Server.app) and configured ** (You cannot use the apple store to install this since it requires the latest version of iOS) * /Deploy smb/AFP SharePoint added to File Share service. dsadmin user given r+w access * Generated a new Netboot set (/Library/NetBoot/NetBootSP0/DSR-10105.nbi) * Enabled and configured NetInstall service * Copied /Deploy from MDC1 install to MDC2 install * Cherry picked workflows, files, scripts, packages and HFS disk images from MDC1 /Deploy to MDC2 Notes: VNC flows are not open from the releng jump hosts into MDC2. Had to flush and disable pf rules in order to connect via VNC. Infoblox DHCP scope for MDC2 test network is not configured to provide DNS parameters to clients. I bypassed this by manually configuring DNS on the OS. Once mdc2 puppet is up and running, this host will need to be re-puppetized with a new cert.
After working with :van, we were able to move the ball forward a little more. TLDR; successfully imaged a mini, puppet needs to be set up to continue the full provisioning. The first issue that wedged the test DC client was a DHCP issue that :rtucker had fix early that morning. DNS was not be set by DHCP and therefore couldn't resolve the DS server fqdn. After that, I was running into an issue where smb_mount would fail with an auth error. This was similar to the issue that we say on 2 minis in MDC1 that we were never able to image because of this. I ended up changing the DS config to use AFP as opposed to SMB/CIFS and ... THIS FIXED IT. I'm going to change this over in MDC1 and see if :van can reimage those 2 orphaned mini from bug 1409857
At this time, deploystuio is functioning and able to re-image mdc2 (pending hardware being racked and powered). The work left to do, which doesn't inhibit deployment, is: * install and configure bacula * setup external SSD and setup time machine * verify reimaging works between test and srv vlan This all depends on the hardware being racked, powered on and accessible in MDC2
Depends on: 1443270
These have all been completed: * install and configure bacula * setup external SSD and setup time machine * verify reimaging works between test and srv vlan
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.