DocScrutinizer05 | anyway it seems to me the box behind http://tableteer.nokia.com is dead. A very usual thing with nokia boxes that are not maintained anymore but still akamai receives payment for the iron | 01:07 |
---|---|---|
bencoh | hmm I dont think akamai is still hosting (or receiving anything) for d.m.n.c | 01:08 |
bencoh | no I mean, even when hardcoding the akamai ip in /etc/hosts, they dont host anything there anymore | 01:08 |
bencoh | so nokia notified akamai | 01:09 |
bencoh | using the old (akamai) ip address (23.205.83.189) doesnt give anything :/ | 16:00 |
---|
Agge | DocScrutinizer05, A friend of min that also ordered a N900 got his today. https://fbcdn-sphotos-d-a.akamaihd.net/hphotos-ak-ash3/t1/q79/s720x720/1969236_752075571478616_608418979_n.jpg | 22:52 |
---|
DocScrutinizer05 | Munrek: watercooler gossip has it that Nemein received order from Nokia to take down the tablets-dev site due to "security problems detected on that site". So the stage site on tablets-dev.maemo.org and the content distribution mirror farm of akamai servers that been tabletsdev.noia.com gone down with a cryptic error msg. When migration completed a year ago, your humble techstaff team found a service on maemo.org disabled and did their | 15:41 |
---|
DocScrutinizer05 | (except that unspeakable stage thing for the nokia.com akamai server farm. But that's stopped since long already) | 03:30 |
---|
DocScrutinizer05 | the akamai servers sometimes have corrupted cache | 12:31 |
---|---|---|
psycho_oreos | ~lart akamai servers | 12:31 |
* infobot whips out a shotgun, trudges over to akamai servers, and goes postal | 12:31 | |
DocScrutinizer05 | seems recent firefox again introduced that "protection" that causes download of stuff from akamai to fail with "double content disposition" or whatever | 12:33 |
psycho_oreos | I recall having to run through that entire process with you awhile ago. Simply using specified akamai's mirror (via DNS or whatever it was) it worked. I'm not going to bother fixing that up for getting just one file. lol | 12:34 |
psycho_oreos | I'll need to also keep a mental note of that, bloody akamai servers. | 12:35 |
DocScrutinizer05 | nobody is servicing those akamai servers anymore | 12:41 |
psycho_oreos | It'll also be interesting to see akamai will work based on IP addresses. There's no way to view the site in "file mode" and it works transparently. | 12:45 |
dos1 | freemangordon: The certificate is only valid for the following names: *.akamaihd.net , *.akamaihd-staging.net , a248.e.akamai.net | 20:15 |
---|---|---|
DocScrutinizer05 | akamai server farm | 20:21 |
DocScrutinizer05 | it's akamai doing the maintenance, and Nokia "forgot" about that infra | 20:24 |
dos11 | but have you seen the true story behind it? https://fbcdn-sphotos-f-a.akamaihd.net/hphotos-ak-ash3/q71/1170764_623665574332364_1926310099_n.jpg | 18:55 |
---|
DocScrutinizer05 | freemangordon: and even when using IP, you still will run into random behaviour of the akamai server farm boxen. We seen this just so often on maemo repos | 23:31 |
---|
Apic | https://fbcdn-sphotos-f-a.akamaihd.net/hphotos-ak-ash4/255018_10152027777808272_2089311588_n.jpg | 11:48 |
---|
DocScrutinizer05 | I won't abandon/abuse a akamai server farm to redirect to a 50EUR/month server at Hetzner ;-D | 15:19 |
---|---|---|
DocScrutinizer05 | psycho_oreos: you might be connected to a akamai server that has garbled content, maybe? | 17:24 |
DocScrutinizer05 | tablets-dev.nokia.com.edgesuite.net is an alias for a1341.g.akamai.net. | 17:25 |
DocScrutinizer05 | a1341.g.akamai.net has address 84.53.175.163 | 17:25 |
DocScrutinizer05 | a1341.g.akamai.net has address 84.53.175.153 | 17:25 |
psycho_oreos | DocScrutinizer05, Similar output: tablets-dev.nokia.com is an alias for tablets-dev.nokia.com.edgesuite.net. | tablets-dev.nokia.com.edgesuite.net is an alias for a1341.g.akamai.net. | a1341.g.akamai.net has address 125.56.218.8 | a1341.g.akamai.net has address 125.56.218.27 | 17:26 |
DocScrutinizer05 | [2013-07-04 16:24:16] <DocScrutinizer05> psycho_oreos: you might be connected to a akamai server that has garbled content, maybe? | 17:27 |
psycho_oreos | I guess it looks like the Google's public DNS is messed up. "dig tablets-dev.nokia.com" outputs "a1341.g.akamai.net. 19 IN A 165.254.99.73 | a1341.g.akamai.net. 19 IN A 165.254.99.33 | ;; Query time: 495 msec | ;; SERVER: 8.8.8.8#53(8.8.8.8)" | 17:30 |
DocScrutinizer05 | akamai runs server farms for content distribution | 17:31 |
DocScrutinizer05 | btw not exactly unheard that parts of the akamai server farm have garbled data | 17:53 |
DocScrutinizer05 | we seen that a lot when repos were on akamai | 17:54 |
DocScrutinizer05 | so you say we traced it down to "akamai has garbled content"? | 18:01 |
psycho_oreos | That's what I would be guessing at this stage by now yes. The garbled content from akamai got me going around in circles. | 18:02 |
DocScrutinizer05 | not much _you_ can do about it. _I_ might try to look into akamai syncing to stage and if sth is fishy in there | 18:03 |
* psycho_oreos wonders if that should be added into the wiki page for future reference. That akamai can give out garbled data. | 18:04 | |
psycho_oreos | At least the mirror link isn't some akamai mirrored link :) which is a :thumbs up: for me. | 18:13 |
psycho_oreos | Seems like the akamai aliases rotate fairly frequently too. Just noticed that its now a completely different set of IP: 69.192.4.67 69.192.4.34 | 18:15 |
DocScrutinizer05 | I bet they got this the same way I do: deploy a private key to repo servers (akamai) | 16:26 |
---|---|---|
DocScrutinizer05 | [2013-05-31 13:14:43] <DocScrutinizer05> <------><------>// Akamai sets this, let's not set it twice | 21:21 |
DocScrutinizer05 | I either comment it out then akamai/nokia.com works, or I don't comment it out then tabletsdev.maemo.org works | 21:23 |
DocScrutinizer05 | <------><------>// Akamai sets this, let's not set it twice | 14:14 |
---|---|---|
jon_y | oh, akamai problem | 14:15 |
DocScrutinizer05 | hosted by maemo.org. Delivered to you by courtesy akamai | 14:18 |
DocScrutinizer05 | (the m.o/tabletsdev content been stage for akamai server farm under tablets-dev.nokia.com) | 14:36 |
---|
DocScrutinizer51 | akamai serverfarm tryng to sync repo to stage | 20:35 |
---|
DocScrutinizer05 | heck, they had an akamai server farm for repo and even for tablets-dev. Now we're servicing that with a single VM on a 100Mb/s uplink, and we have lots of headroom | 06:12 |
---|
DocScrutinizer05 | ecc3g: nope, since that been Akamai server farm | 03:18 |
---|
DocScrutinizer05 | plus a fancy little server farm for repo at Akamai | 22:14 |
---|
RST38h | https://fbcdn-sphotos-c-a.akamaihd.net/hphotos-ak-prn1/48127_10151401467333908_1480734954_n.jpg | 18:33 |
---|---|---|
M4rtinK | BTW, the akamai statistics showed about 40k N900s | 23:33 |
DocScrutinizer05 | repo moved from a Akamai server farm with several hundred servers to a VM with a 50Mbit uplink | 09:31 |
---|---|---|
pwnguin | akamai server farm seems like overkill | 09:31 |
DocScrutinizer05 | LinuxCode: X-Fade mentioned there are some 30k to 40k active, according to Akamai repo server farm statistics | 02:21 |
---|
kerio | it's not easy to keep a server up when akamai DDoSes it | 20:58 |
---|---|---|
kerio | because nokia used akamai | 20:59 |
kerio | and when the migration happened, akamai decided that it had to redownload the whole of rmo | 20:59 |
LinuxCode | DocScrutinizer05, if akamai keeps hammering | 21:10 |
DocScrutinizer05 | akamai servers are down, DNS stage.maemo.org is gone | 21:11 |
DocScrutinizer05 | qwazix: small update, we came from some 20 physical servers plus Akamai server farm, the current config already is massively consolidated by Nemein | 00:15 |
---|---|---|
kerio | is akamai still all up in our grill? | 00:55 |
DocScrutinizer05 | akamai serverfarm seems to have shut down. AIUI it's same servers that host tablets-dev | 00:59 |
ShadowJK | points to akamai | 13:57 |
---|---|---|
DocScrutinizer05 | and those akamai are still existent? | 13:58 |
kerio | DocScrutinizer05: why is it so slow? still akamai? | 13:33 |
---|---|---|
Pali | I think both are akamai | 23:09 |
kerio | but it's still akamai | 23:09 |
warfare | Pali: That's the error when akamai can't reach its origin servers. | 23:14 |
DocScrutinizer05 | we got a DNS for repository.maemo.org already pointing to IP of the VM originally assigned for that purpose. We found out that akamai polling stage.maemo.org instantly killed the rack's uplink. We took down stage.m.o VM and things somewhat became normal, but Eero realized that actaully stage.m.o is our new *real* repository.m.o. Eero implemented a firewall redirection from repo.m.o IP to stage.m.o IP/VM. When we started up stage.m.o VM | 02:08 |
---|---|---|
DocScrutinizer05 | again, it again instantly killed *.m.o by congestion of some bottleneck in uplink. At that stage we shut down stage.m.o again and called it a week on Friday afternoon. Monday we will investigate whether maybe the akamai->stage.m.o path still was open and thus starting of stage.m.o VM caused same issue we initially seen on Wednesday | 02:08 |
DocScrutinizer05 | when vanishing, might fix akamai DDoS attack for good | 02:36 |
ShadowJK | and now it was just shut down waiting for akamai to stop | 00:01 |
---|---|---|
qwazix | wtf is akamai? | 00:20 |
Ultima_Fan | akamai, the keeper of the prophecy! | 00:21 |
DocScrutinizer05 | qwazix: repository.maemo.org pointed to akamai server farms | 00:23 |
DocScrutinizer05 | since akamai has/had *plenty* of those servers trying to sync the repos from stage.m.o | 00:26 |
DocScrutinizer05 | repository.maemo.org.edgesuite.net. 21600 IN CNAME a515.g.akamai.net. | 00:34 |
DocScrutinizer05 | a515.g.akamai.net in turn will resolve to different IP depending on your geolocation | 00:35 |
DocScrutinizer05 | actually I guess a515.g.akamai.net is ~300 servers worldwide | 00:35 |
ShadowJK | it's turned off on purpose waiting for nokia to fix dns.. there was a problem with akamai hogging all bandwidth trying to mirror everything | 16:42 |
freemangordon | chem|st: we wait for akamai to fix something? | 18:01 |
freemangordon | well aiui akamai did a DOS attack on the repos :D | 18:02 |
kerio | DocScrutinizer05: points to akamai | 18:38 |
DocScrutinizer05 | so probably Nokia taking down akamai servers right now | 18:39 |
kerio | well, it's kinda hard to keep servers up when you're being DoSed by akamai | 10:52 |
---|---|---|
tommis | akamai? | 10:52 |
* DocScrutinizer51 idly ponders to turn those akamai server farms into a nice botnet, since probably Nokia will pay for them another 2 years :-P | 11:25 |
DocScrutinizer05 | xes: a short history: *.maemo.org lived on ~16...20 akamai(?) server(-clusters) owned by Nokia, together with maemo-unrelated stuff. Nokia decided to "donate" all maemo stuff to community and ordered Nemein (their hoster/maintenance service) to move stuff to dedicated locations that have no unrelated bits anymore, and then donate that to community | 00:06 |
---|
DocScrutinizer05 | repo is with akamain still | 11:13 |
---|---|---|
wirr | repository.maemo.org is on akamai? | 11:13 |
DocScrutinizer05 | repository.maemo.org.edgesuite.net. 21600 IN CNAME a515.g.akamai.net | 11:14 |
DocScrutinizer05 | [2013-01-14 10:14:28] <DocScrutinizer05> repository.maemo.org.edgesuite.net. 21600 IN CNAME a515.g.akamai.net | 16:02 |
DocScrutinizer05 | i'd not be surprised when Nokia or akamai takes down the 'old' servers same day they started to change DNs to new ones | 16:03 |
DocScrutinizer05 | yeah, the akamai servers didn't work this morning either | 19:44 |
DocScrutinizer05 | kudos for the Akamai servers' backbone though | 15:56 |
---|
Raimu | OT: http://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-snc7/s720x720/392380_3645934780207_1029526329_3281305_2046772520_n.jpg these EULAs are getting worse all the time. | 15:46 |
---|
k273 | https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-ash4/385164_245474242178126_153853541340197_698259_1028326375_n.jpg | 05:45 |
---|
100 matches in 3289 log files with 1386123 lines (4.6 seconds).
Generated by irclogsearch.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!