As some of you already might have noticed, this forum as well as all *.maemo.org infra is going to move to a new home, due to Nokia not providing/supporting any hosting on Nokia servers for it anymore (well, for tmo the case is slightly different, but that's irrelevant here).

Anyway
Wednesday 2013-01-16 will be the magic date, Reggie just informed me that the vBulletin incl database got copied to new server and tested successfully, and on Wednesday we'll
put this server into r/o mode and again copy over all data to the new location.
Let's hope we got control over the nameservers of maemo.org until then, so we can point the URL to the new server same date. Otherwise check http://mwkn.net/2013/01/community.html
and more recent posts there for instructions how to access the new servers during the times when we still sort out DNS issues.

Please don't forget that keeping this new infra running for you is only possible when Hildon Foundation can pay for the new servers and warrant maintenance, so please read http://talk.maemo.org/showthread.php?t=88222.
If you're a professional sysop/admin used to taking care of infrastructure like 12 VM on 2 octocore 64GB dedi servers on a regular basis during your daywork, and you feel like helping maemo.org with the bare system maintenance (security updates, monitoring, etc) on a dumping price level, please mail council@maemo.org subj: "sysop" (ML are temporarily out of order due to migration, so please CC joerg[at]openmoko.org)

sincerely
jOERG
(maemo community councilor, maemo.org administration coordinator)

[update]
Nokia dnsmaster been asked to do the DNS switch at Wed 10:00 UTC. Since some of the other domains already got moved (e.g. wiki is back to normal, HOORAY!), odds are this might just work in time :-)


[update2 2013-01-14_20:25UTC]
Obviously almost all DNS records have been updated to new IP addr during this day already, and tmo A record got replaced by a CNAME (whatever the rationale behind that). The effect for now is that the new servers have exposed some network issues which cause extreme slowdown or even downtime for *.maemo.org - except talk.maemo.org. Nemein is working to fix stuff. By no means you should run any mirroring, wget, or other bandwidth-hogs against *.maemo.org right now, since the interim infra might not be able to handle that.
Thanks for your attention.


[update3 2013-01-15_15:00UTC]
[[#maemo-meeting]]
<bergie> seems there is an issue with the firewall, I'm talking with our support guys ATM
<bergie> but apparently there is something in routing that needs to be looked at
<bergie> yep, the firewall is shutting down quite a few of the connections. We're talking with the ISP
<bergie> I'll let you know more as soon as I have the info
<DocScrutinizer05> thanks
(see http://mg.pov.lt/maemo-meeting-irclo...01-15.log.html)


[update4 2013-01-15_23:40UTC]
[2013-01-15 19:13:11] <mashiara> I got some more info but not enough to make any better guesses at the problem than previously. I will hopefully get some real data tomorrow during office hours.
[2013-01-15 19:13:36] <mashiara> I'm going to call it a night now.
[2013-01-15 19:27:42] <DocScrutinizer51> ok, so no real change in situation to be expected until <now>+12h
[2013-01-15 19:29:11] <DocScrutinizer51> IOW tomorrow 0700UTC
[2013-01-15 19:30:15] <DocScrutinizer51> my latest tests seem to indicate that repository and maemo.org are available in one out of 10 tries

Due to the persisting issues with *.maemo.org I asked Reggie to postpone tmo migration (or rather the DNS switching) for a day or two, until the issues with connectivity are solved. This might mean the server will go into r/o mode as announced above, but stay in r/o mode until aforementioned problem got solved, and only then we switch DNS to new server, a day or two later.

[update5 2013-01-16_19:55UTC]
we finally got the major issues sorted! repository.maemo.org still down, but that has no impact on tmo migration. Reggie is about to initiate that right now.
Many thanks Reggie! :-) And a heartly welcome and thanks for all the help during migration work so far to our new designated master admin of tmo: chemist. He'll take over from Reggie after migration in a relaxed cross-fade process of responsibilities. :-)
[[update-ps]] the repository.maemo.org is waiting for DNS change to point to stage.maemo.org vserver. Eero decided to implement this migration detail in this way, so we wait for Nokia to edit the zone-files on nameserver.

[update6 2013-01-18_15:15UTC]
MISSION ACCOMPLISHED Talk.maemo.org moved and active. Many thanks Reggie! :-)
General maemo.org migration ongoing.

[update7 2013-01-19_15:30UTC]
tmo is missing some config and support form other parts of m.o-infra to send emails. This will not get sorted before Monday.
For the repositories: they are still down since we are still facing bandwidth problems that took out complete *.maemo.org infra yesterday the very moment we enabled repository.maemo.org via a Firewall redirect. Despite there being some 40,000 devices out there trying to auto-update once a day, this doesn't exactly explain the problems we see. We are investigating the issue and hope to bring back standard repositories eventually. For those in a real pinch: please read http://talk.maemo.org/showthread.php?t=88707 for a community driven temporary workaround, that may or may not work for you

[update8 2013-01-21_01:15UTC] (sorry, no real news since update7)
[edited, plans changed] mail from tmo fails [del]. Ruediger and Reggie will manage direct SMTP inside forum software [del]. Nothing users of tmo could do to help with this. If you want to help with migration, please configure your N900 to NOT do daily automatic updates, so load from 40000 N900 trying to update on repository.maemo.org gets lower.

[update9 2013-01-22_08:30UTC]
tmo mail supposed to mostly work again, some issues with *your* spamfilter might get sorted during the day by chemist. If you can't receive notifications from tmo, then changing your mail addr in tmo to point to something with less strict spam filtering might help.
repo: alas no news:
[2013-01-21 15:29:54] <DocScrutinizer05> mashiara: any news about bottleneck?
[2013-01-21 15:30:18] <mashiara> no, but I haven't pestered the ISP about it today either
[2013-01-21 15:30:40] <mashiara> I did pester Nokia about the DNS thing but have gotten no replies yet
[2013-01-21 15:32:34] <mashiara> anyway I'm not sure if I've mentioned it but my coworker broke his leg on new-year and is on sick leave until end of february, so I'm more than a little swamped with work here.
((your humble admin coordinator and author of these lines also got swamped with RL issues recently, so please bear with me when there's an occasional delay on pushing migrations and updates about it))

[update10 2013-01-23_00:30UTC]
This morning (14h ago) we faced another slowdown of services. Firewall been taken down, a CPU got added, and everything started again. Since then things seem incredibly fast in comaprison to before. We planned to give stage.m.o aka repository another try but prior to that more tweaks to firewall will be needed and we didn't finish that today. However if everything pans out as expected, odds are we might bring back repo tomorrow.

[update11 2013-01-23_12:00UTC]
Some tuning on FW got done. 30 minutes ago we enabled repository.m.o VM. Heavy traffic drives the connection to its limit and causes extreme load times or even connection rejects. The rest of infra nevertheless seems to stay stable. We'll watch the situation and see how long it takes til those hungry 40000 devices out there slowly decay their DDoS attack.

[update12 2013-01-24_08:55UTC]
When yesterday we enabled stage.maemo.org (aka repository.maemo.org) we seen another crash of FW after 90min. Reboot. Gave it a 2nd chance. This morning (2h ago) I found *.m.o down again. We took down the stage VM and gave FW another reboot. So current state: everything "fine", except repository.maemo.org. Current plan: expose repository.maemo.org to the internet via bypassing this FW POS. We're waiting for the "grid-guy" to establish that. Sorry for the inconvenience.
[update12a 2013-01-24_15:55UTC] garage.maemo.org VM has problems, is down, and authentication went down with it. *.maemo.org is semi-ok as long as you don't try to authenticate. ETA: 90min

[update13 2013-01-26_03:30UTC]
I don't believe in such stuff, but update13 seems the right name here anyway: since my last update we seen further outages of FW, database (unconfirmed, indication been 503 on maemo.org), and FW again and again. Currently everything came to a grinding halt again (as usual except this forum, which - despite on same rack - doesn't run thru the FW). Alas my access to that FW is still limited (actually non-existent) so I hope for Eero to at least give it another kick in a few hours, before he leaves for weekend. Sorry for the inconvenience. On the bright side: Ferenc fixed a few bugs in garage, so stuff should work better there now. Pali already said he been able to check in stuff for the first time since ages :-). Well, as long as you can reach it...

[update14 2013-01-29_11:45UTC]
Sorry for "long time" no update, I got a temporary nervous personal downtime. After actually suffering a *.m.o downtime during the weekend, FW finally seen reboot and slight reconfigs on Sunday evening, thanks Eero for that. Since then we're mostly back to our semi-operational state: everything but repository reachable, problems with login on maemo.org top page (not though on bugs, garage, tmo, wiki). Exposing repo.m.o to bypass FW is WIP. Announcing Falk Stern alias warfare as our first approved maemo sysop volunteer. Bug in tmo search function got fixed by re-indexing, courtesy Ruediger.

[update14 2013-01-31_23:45UTC]
repository.maemo.org enabled and kinda stable, though heavy load slows down stuff for the particular client. HAM catalog update might take up to 1h..

[final update 2013-02-17]
repository.maemo.org is technically working. We're still suffering legacy bugs that cause hashsum errors - also in most mirrors, though depending on when they did last mirroring it seems to differ between them. Merlin's mirror however is a redone repo, not a true mirror, and thus considered free of hashsum bugs.
We also have no autobuilder yet, and login on some subsystems doesn't work. WIP

[links]
http://wiki.maemo.org/Migrating_to_C...Infrastructure

Thank u a lot guys for this great work .. u are great people .. Maemo will die without your effort . .

Yes indeed, thanks guys for your outstanding work and dedication to us Maemo lovers!

I saw your heated debate about whether the DNS for TMO temporarily should be moved as well.
I'm glad you didn't do that, even if many of us are not grateful enough and unaware of the hard work going on behind the scenes its still our most important waterhole for information about our beloved devices.

Good thing you guys didn't loose your temper and head away with name calling and such.

As a first post, many congrats on what you did. I came here to sort this problem out; thank god i used the search function first. Can't wait to try my new N900 out!

Great job and greetings from Belgium

Nice work and congrats chemist!

I know the servers are down and i dont want any hate posts for this but

as maemo.org seems to be back up does anyone know when the repos will be working.. ive just got a new n900 but cant seem to update repos im sure this is because the servers were being migrated which is fine but all i need to know is this why the repos dont work?

never mind i read the last update. Move along Lucas nothing to see here.

Just letting you all know that I will start the migration in a few minutes. I will be doing the following:

• forum will be turned-off
• create a backup TMO
• re-open the forum in read-only mode
• restore the backup to the new server
• wait for DNS change to take effect

Keep checking the site. Once you notice that people can post again, then the move is done.

Thanks.

Alelujia !

Please report any bugs or problems you encounter.

Thanks!