Closed Bug 740613 Opened 14 years ago Closed 14 years ago

four new VMs in scl3 for seamonkey

Categories

(Infrastructure & Operations :: Virtualization, task)

Product:
Infrastructure & Operations
Component:
Virtualization
Type:
task
Priority:
Not set
Severity:
major

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: dustin, Assigned: afernandez)

References

Details

Seamonkey's VMs are moving - mostly from parallels - from sjc1, so we'll need the following. All are in community.scl3.mozilla.com. sea-master1 centos6 or rhel6 63.245.223.11 2GB RAM 30GB disk not puppetized - I will set it up for Callek sea-vm-linux64-1 based on the templates and specs used in bug 735381 I or releng will clean it out sea-vm-linux32-1 migrated from cb-sea-linux-tbox (ESX) in sjc1 sea-vm-win32-1 migrated from cb-sea-win32-tbox (ESX) in sjc1 The hosts below will be cloned from similarly-named hosts above, once they're configured. I think that requires no action from the virtualization team right now, but they're here as an FYI. sea-vm-linux32-2 sea-vm-linux32-3 sea-vm-linux32-4 sea-vm-linux32-5 sea-vm-linux32-6 (later, when the .ams1 hosts are turned off) sea-vm-win32-2 sea-vm-win32-3 sea-vm-win32-4 (later, when the .ams1 hosts are turned off) This will need to be done well in advance of 4/23 - let's say 4/16. So it can happen after the ongoing mad-dash to 4/9.
I'll have the IPs in DNS by the time you act on this - there's some changes pending to the allocation.
(In reply to Dustin J. Mitchell [:dustin] from comment #0) > Seamonkey's VMs are moving - mostly from parallels - from sjc1, so we'll > need the following. All are in community.scl3.mozilla.com. > > sea-master1 > centos6 or rhel6 > 63.245.223.11 > 2GB RAM > 30GB disk > not puppetized - I will set it up for Callek Hrm really? Is that the OS Releng uses for masters right now? I thought they had cent5.5 and my current one has cent5.0 (because I was confused a bit here). If they do have cent6 currently, I'm all for it, and I do note that my current master is linux32, relengs are certainly linux64, 64 is probably a better choice.
(In reply to Dustin J. Mitchell [:dustin] from comment #0) > The hosts below will be cloned from similarly-named hosts above, once > they're configured. I think that requires no action from the virtualization > team right now, but they're here as an FYI. Also to be clear, for moving the parallels-based hosts off that system I'd want to build/create these new hosts fresh (or from a releng image with keys cleared, etc.) Then we can overwrite the ones we are moving from -tbox. But we do have the parallels host scheduled to move anyway, though if I get all the VMs up before that move we don't even have to move that machine.
Revising comment 0, then: Seamonkey's VMs are moving - mostly from parallels - from sjc1, so we'll need the following. All are in community.scl3.mozilla.com. sea-master1 centos5.5 32-bit (to replicate releng systems) 63.245.223.11 2GB RAM 50GB disk (my math was wrong in comment 0, plus room for growth) basic OS install - I can take it from there sea-vm-linux64-1 based on the templates and specs used in bug 735381 I or releng will clean it out sea-vm-linux32-1 migrated from cb-sea-linux-tbox (ESX) in sjc1 sea-vm-linux32-2 (rather than cloning the ancient -tbox image) based on the templates and specs used in bug 735381 I or releng will clean it out sea-vm-win32-1 migrated from cb-sea-win32-tbox (ESX) in sjc1 sea-vm-win32-2 win2k3sp1 either from the old releng win32 image in sjc1, or a fresh OS install, whichever is easiest The hosts below will be cloned from similarly-named hosts above, once they're configured. I think that requires no action from the virtualization team right now, but they're here as an FYI. sea-vm-linux32-3 sea-vm-linux32-4 sea-vm-linux32-5 sea-vm-linux32-6 (later, when the .ams1 hosts are turned off) sea-vm-win32-3 sea-vm-win32-4 (later, when the .ams1 hosts are turned off) This will need to be done well in advance of 4/23 - let's say 4/16. So it can happen after the ongoing mad-dash to 4/9.
Summary: three new VMs in scl3 for seamonkey → four new VMs in scl3 for seamonkey
I will be learning how to do this by doing it. Someone has to :) The network isn't re-allocated yet, so this isn't quite ready to go yet.
Assignee: server-ops → dustin
Matt, can you work on this? It's slipping down on my todo list :(
Assignee: dustin → mlarrain
aj can take this.
Assignee: mlarrain → afernandez
Depends on: 721516
This is a big list of hosts, and Callek's going to have a lot of work to do to stand them up before we can take down the set in sjc1. When can this fit into the schedule?
Severity: normal → major
Should have all VMs up today. Will contact you via IRC if any questions pop-up to speed up resolution.
Status: NEW → ASSIGNED
Got delayed, should have ALL done by morning-noon. (aiming for morning)
Delayed again, working on this and will update when done (today!)
sea-master1.community.scl3.mozilla.com is online sea-vm-linux64-1 and sea-vm-linux32-2 were deployed from the specified templates from bug 735381 which are using puppet and dhcp IPs (these have to be set statically), however, there might be issues with the templates (VMs are up but not online). sea-vm-linux32-1 and sea-vm-win32-1 will migrated soon. As for sea-vm-win32-2 will need to be install from scratch (can't find template).
Depends on: 747476
Aj, sea-vm-linux64-1 and sea-vm-linux32-2 are on the wrong VLAN. sea-master1 appears to be on the correct VLAN, but doesn't have VMWare tools installed, and isn't pingable at its correct address. Can you fix that up? Thanks!
Aj? Any progress here?
sea-master1.community.scl3.mozilla.com is online. The reason why sea-vm-linux64-1 and sea-vm-linux32-2 are on different vlans is for testing purposes. Seems that the template(s) from Bug 735381 may be to blame. :Callek are you aware that the template(s) from Bug 735381 are puppetized? Will continue looking into it tomorrow and will also continue the migration(s) tomorrow.
Aj, I'll clean those puppetized VMs out. Are they up now? I don't see them in the community VLAN.
Dustin, VMs are online now; sea-vm-linux64-1 sea-vm-linux32-2 I manually set the network configuration via the cli (seems the networking scripts are set-up to only work via puppet). Working on the other vms.
Migration of sea-win32-tbox started; 1% completed ETA: ~ 1hr (should be less)
For general reference, the ETA could fluctuate drastically (go up/down).
sea-win32-tbox migration completed. New host name: sea-vm-win32-1.community.scl3.mozilla.com cb-sea-linux-tbox next.
cb-sea-linux-tbox migration in progress. Status: 1% Complete ETA: ~ 1 hour and 22 minutes. (will fluctuate)
sea-vm-linux64-1 cleaned per releng instructions passwords reset to seamonkey defaults cltbld deleted; seabld created ready for callek sea-vm-linux32-2 cleaned per releng instructions passwords reset to seamonkey defaults cltbld deleted; seabld created ready for callek sea-vm-win32-1 not at its assigned IP I don't have vmware access to investigate sea-master1 ready for callek
cb-sea-linux-tbox migration status; Status: 84% Complete ETA: ~ 20 minutes. (will fluctuate) Main reason for long delay, some congestion on the network.
Migration completed, getting server online.
cb-sea-linux-tbox now known as; sea-vm-linux32-1.community.scl3.mozilla.com is online. As for sea-vm-win32-2 should be done tonight/tomorrow. Currently uploading the VM from bug 747476 into scl3.
Forgot to add, regarding sea-vm-win32-1.community.scl3.mozilla.com it's up. :dustin verified as well.
(In reply to Dustin J. Mitchell [:dustin] from comment #0) > sea-master1 ready to go (root acct only, SM password) > sea-vm-linux64-1 ready to go > sea-vm-linux32-1 up and running > sea-vm-win32-1 ready to go So, I think this is over to Callek to verify and get things ready to clone as described in comment 0. Callek, please open a new bug for the clones, when you're ready.
Assignee: afernandez → bugspam.Callek
Whoops, I should have looked at comment 4. That just adds one host: (In reply to Dustin J. Mitchell [:dustin] from comment #4) > sea-vm-win32-2 still in progress from comment 25; I'll need to clean this once it's cloned, so back to Aj for the moment :)
Assignee: bugspam.Callek → afernandez
sea-vm-win32-2.community.scl3.mozilla.com VM is up but cannot login as Administrator. The default (automatic login) account does not have Administrative privileges. The template is based off win2k3sp2-vc8tools-ref-vm Looking for the password.
ok found password (RelEng one). VM sea-vm-win32-2.community.scl3.mozilla.com is now online. The specs (looking at other bug:735381 ) are as follows; 100GB disk 2 CPU 4 GB RAM However, the original template (win2k3sp2-vc8tools-ref-vm) had 2 hard drives, I only increased the 1st (OS) hard drive to 100GB and left the 2nd hard drive as is. Please let me know if the 2nd hard drive should be removed. If this is incorrect, please let me know. Please verify that all is well with this one and ALL other VMs.
OK, host is cleaned. I deleted cltbld, and the Administrator pw is the usual. I'll leave it to you (Callek) to create seabld. I also renamed the host. Aj, go ahead and remove the second disk. If that doesn't involve a reboot, please do a reboot to get the new hostname set up.
ok sea-vm-win32-2.community.scl3.mozilla.com is 100% done, the template VM is called: sea-vm-win2k3sp2 Please verify that the VM is good so that the template could be used in bug 748476.
(In reply to Dustin J. Mitchell [:dustin] from comment #16) > Aj, I'll clean those puppetized VMs out. Are they up now? I don't see them > in the community VLAN. Were these run through [current] puppetization or am I in a state where I need to manually apply the missing rpms/files/etc. that puppet would have? (In reply to Adrian Fernandez [:Aj] from comment #30) > ok found password (RelEng one). > > VM sea-vm-win32-2.community.scl3.mozilla.com is now online. > The specs (looking at other bug:735381 ) are as follows; > 100GB disk > 2 CPU > 4 GB RAM > > However, the original template (win2k3sp2-vc8tools-ref-vm) Was there a ref named "win2k3sp2-ref-vm"? which was the memory of a relenger if not then we can be good > had 2 hard > drives, I only increased the 1st (OS) hard drive to 100GB and left the 2nd > hard drive as is. > Please let me know if the 2nd hard drive should be removed. > If this is incorrect, please let me know. > > Please verify that all is well with this one and ALL other VMs. (In reply to Dustin J. Mitchell [:dustin] from comment #31) > Aj, go ahead and remove the second disk. If that doesn't involve a reboot, > please do a reboot to get the new hostname set up. Actually the newer Releng VMs that are Talos based only use 1 drive, deleting the D:\ drive was a wrong move (lost some of the moz-build/etc. installs) and we would have wanted to add an E:\ (Builds) with the remaining disk space (I'm hoping for ~80GB if the space allows) -- As far as using it for NSS, it would probably be almost identical but need a larger C (by about 10GB) to be able to install MSVC2010, if its better/easier to have me set this up for their basic use as well, lets plan on: * ReClone the VM * Leave cltbld on it (so I can give NSS a system with that user) * 20GB C:\ [expanded from refimage], default D:\ from releng refimage, and 70-80GB E:\ (added fresh) * Releng Passwords cleaned/changed by dustin Does that sound sane/useful/possible?
Blocks: 748476
Clarified some things with :Callek , proceeding with request(s).
New re-cloned sea-vm-win32-2.community.scl3.mozilla.com is up again. :dustin please clean what needs to be cleaned. :Callek once "cleaned" please verify that all is well so that i could then clone the VM for 748476
cleaned.
..and passwords are changed. Per request, I did not delete the cltbld account.
:Callek any progress?
(In reply to Adrian Fernandez [:Aj] from comment #38) > :Callek any progress? For clarity, I have this ready for NSS/Calendar cloning in seperate bugs, once those are done I need to swap the user over to seabld instead of cltbld.
:Callek you should be good to go on sea-vm-win32-2.community.scl3.mozilla.com (ie make any further changes). sea-vm-win32-2.community.scl3.mozilla.com was cloned as template: win32-build (if you have a better name, please let me know). Essentially this particular bug could be closed now?
Verified with :Callek that as far as this bug is concerned, all is done.
Status: ASSIGNED → RESOLVED
Closed: 14 years ago
Resolution: --- → FIXED
Blocks: 764693
Blocks: 773147
Product: mozilla.org → Infrastructure & Operations
You need to log in before you can comment on or make changes to this bug.