In the last few hours, two different mirror admins have reported to #mirrors that their traffic levels have dropped off dramatically. I checked my mirror's levels, too, and I'm seeing the same thing. I'm still getting some traffic, but it's ridiculously low (only 130.68Mbps at weight 25000). Something is wrong...
This should be back - I was testing something and added it with a mirror weight of 50,000. Are you still seeing depressed traffic?
(In reply to comment #1) > Are you still seeing depressed traffic? Yeah, I'm down to 121.96Mbps right now.
What are you expecting to see? I topped out at just under 2Gbps earlier which matches what I've seen in the past on a Friday past peak during a non-release.
(In reply to comment #3) > What are you expecting to see? I topped out at just under 2Gbps earlier which > matches what I've seen in the past on a Friday past peak during a non-release. Generally, 1-1.5Gbps on average.
I'm still seeing a bunch of traffic to this Ankeena box even though it's marked as disabled (and no longer showing up on the sentry page).
I've been watching one webhead for about 8 min now and here are the top 10 mirrors the app is selecting: 956, http://www.mirrorservice.org 1019, http://pv-mirror01.mozilla.org 1109, http://mozilla2.snt.utwente.nl 1147, http://mirror.leaseweb.com 1159, http://mirror.yellowfiber.net 1198, http://mozilla3.snt.utwente.nl 1205, http://ftp.sunet.se 1409, http://hyperion.zih.tu-dresden.de 2255, http://ftp.jaist.ac.jp 2434, http://mozilla.mirror.ac.za
Top 10 after 40 min: 4634, http://www.mirrorservice.org 4755, http://pv-mirror01.mozilla.org 5283, http://mirror.leaseweb.com 5300, http://mozilla2.snt.utwente.nl 5374, http://mirror.yellowfiber.net 5439, http://mozilla3.snt.utwente.nl 5703, http://ftp.sunet.se 6570, http://hyperion.zih.tu-dresden.de 10276, http://ftp.jaist.ac.jp 11185, http://mozilla.mirror.ac.za
So, why am I (the top mirror at weight 25000) not in that list?
12:55 < oremj> mrz: in the database voxcdn is not active for 3.6.3 I did note this earlier: Log entry for  () at 2010-04-30 17:17:29 UTC Note: a FAILED/404 result on a file which is not included in the mozilla-current module is okay if you are only rsyncing mozilla-current. Checking mirror 3347-mozilla.voxcdn.com ... http://3347-mozilla.voxcdn.com/pub/mozilla.org sent no response after 5 seconds! Checking recent history... **** weight 25000 active 0 for http://3347-mozilla.voxcdn.com/pub/mozilla.org **** weight 25000 active 1 for http://3347-mozilla.voxcdn.com/pub/mozilla.org **** weight 25000 active 1 for http://3347-mozilla.voxcdn.com/pub/mozilla.org **** weight 25000 active 1 for http://3347-mozilla.voxcdn.com/pub/mozilla.org Pattern OK, leaving weight unchanged.
(In reply to comment #12) > 12:55 < oremj> mrz: in the database voxcdn is not active for 3.6.3 Uh, why not? I'm pulling mozilla-releases, and I have 3.6.3. http://3347-mozilla.voxcdn.com/pub/mozilla.org/firefox/releases/3.6.3/ > Log entry for  () at 2010-04-30 17:17:29 UTC Looks like it's missing some information there? > Checking mirror 3347-mozilla.voxcdn.com ... > http://3347-mozilla.voxcdn.com/pub/mozilla.org sent no response after 5 > seconds! Dunno... Other mirrors have been having sporadic red blocks as well... but seems everything is stable on my end.
(In reply to comment #11) > s/you/voxel/ :) Eh, Voxel just donates the bandwidth and the CDN network. The origin server that serves the CDN its content is my personal server (for which I personally pay).
Found the problem. It's a bug in sentry.
Fixed in r66666.
Status: NEW → RESOLVED
Last Resolved: 9 years ago
Resolution: --- → FIXED
Quick 411 on the bug: If you run a query with the Perl MySQL interface that returns the column "id" from two different tables, the latter will overwrite the former. Makes me want to abandon Perl sooner rather than later. Fix is here: http://github.com/fwenzel/tuxedo/commit/545dc87 and in r66666.
Product: mozilla.org → mozilla.org Graveyard
You need to log in before you can comment on or make changes to this bug.