You are not logged in.
I've read the thread on what to do when the AUR is down, but I have to make this thread because it's an especially odd issue that may or may not fall in the category of the AUR being down. Through yesterday and today, the AUR has been screwy. How so? Well, multiple AUR helpers (I'm using yaourt and pacaur) are spending 90% of the time acting like there's no connection, but ocassionally getting through, as if there's extreme packet loss. However, when I ping aur.archlinux.org, it seems fine. I tried pinging at the exact same time that yaourt and pacaur were unable to connect, and the ping was fine. No packet loss, no slowdown, nothing. But then I noticed it wasn't pining aur.archlinux.org. Though that was what I had typed in, it was instead "pinging luna.archlinux.org". When I tried browsing to that URL, it redirected me to the Arch Forums. Ontop of that, I haven't just tried pinging. I've also tried browsing to aur.archlinux.org while both AUR helpers couldn't connect, and it loaded perfectly. It wasn't even slow or finnicky about it. So, multiple weird things going on that I don't know enough to comment on, but want fixed.
Oh, and the exact error that gets thrown when the AUR helpers can't connect is, on yaourt:
curl error: SSL connect error
every time.
Last edited by Cadeyrn (2015-12-28 20:48:40)
Offline
Ignoring the AUR helpers....
When you browsed to the aur and got the forums, did you really get the forums, are did you get the AUR home page?
Understand that the AUR is essentially a git repository of build scripts. The AUR home page is a search interface that lets you find the URL of the git repository for any AUR entity.
Once the git is cloned from the AUR, the makepkg script is run against the PKGBUILD script of the AUR entity. The data downloaded from the AUR is usually only several hundred bytes.
Then the heavy lifting begins. The source code is downloaded from wherever it lives on the Internet. One place those data are not is the the Arch AUR.
So, what is taking forever? The AUR, or the download sites? Since you are using helpers, we cannot tell by your description. It is entirely possible that the git half of the AUR was down, but this is the first I've heard of it.
My suggestion, follow the instructions here. Use the instructions for using git instead of tarbels. If things break, provide us the specific commands you used, and all of the program output.
Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way
Online
When I browsed the AUR (aur.archlinux.org), I got to the AUR page. When I browsed luna.archlinux.org, I got to the forums. Both worked fine during the AUR helpers' screwiness, but I mentioned luna.archlinux.org because I've never heard of that URL and don't understand why A) ping would be redirected from aur.archlinux.org to it, and B) why Firefox would be redirected to the forums from it.
It's the AUR itself that will not load in the AUR helpers but will load in my browser. The AUR helpers immediately get stuck and eventually timeout with that error before giving any output whatsoever. Since I can still browse to the AUR site, I'd imagine downloading the PKGBUILDs and installing them manually would also work, but downloading the PKGBUILD is a step that somehow does NOT work through the AUR helpers.
I thought maybe it was an issue with curl that doesn't affect web browsers or ping, but I just tried curl http://aur.archlinux.org/index.html at the same time that yaourt couldn't connect and threw the curl error, and it worked. And... now it's getting weird. OK, forget everything I said. The issue's been too intermittent to get this data until just now.
I tried, once again, both AUR helpers, curl, a web browser, and a ping while the AUR helpers couldn't connect, and the other things also couldn't connect. Nothing could. At the same time, I checked downforeveryoneorjustme and it said it was just me. So, this problem got a lot simpler. The AUR is intermittently down for just me without any other website (including other Arch Linux websites) being down for me. Should I ask my ISP about this?
Last edited by Cadeyrn (2015-12-28 21:24:18)
Offline
A) ping would be redirected from aur.archlinux.org to it, and B) why Firefox would be redirected to the forums from it.
'luna' is the name of the server where both the forums and AUR are hosted, using Virtual Hosting.
I tried, once again, both AUR helpers, curl, a web browser, and a ping while the AUR helpers couldn't connect, and the other things also couldn't connect.
My best guess would be a routing issue with your ISP based on what you've described so far. Post the output from the following commands next time it is failing:
host aur.archlinux.org
ping -nc3 aur.archlinux.org
traceroute aur.archlinux.org
wget https://aur.archlinux.org/
(You may need to install bind-tools and traceroute if you don't already have them installed; ping should be installed as part of base)
Are you familiar with our Forum Rules, and How To Ask Questions The Smart Way?
BlueHackers // fscanary // resticctl
Offline
Alright, I finally got a good run-down of the situation. I really confirmed that it's just the AUR that has the problem, and even archlinux.org works perfectly fine while I can't get to aur.archlinux.org. But, the other commands didn't have many issues. Just Firefox, wget, yaourt, and pacaur were the programs that couldn't access the AUR during the mini-outage.
wget:
$ wget https://aur.archlinux.org
--2015-12-28 14:18:50-- https://aur.archlinux.org/
Resolving aur.archlinux.org (aur.archlinux.org)... 5.9.250.164, 2a01:4f8:160:3033::2
Connecting to aur.archlinux.org (aur.archlinux.org)|5.9.250.164|:443... connected.
HTTP request sent, awaiting response... No data received.
Retrying.
--2015-12-28 14:20:00-- (try: 2) https://aur.archlinux.org/
Connecting to aur.archlinux.org (aur.archlinux.org)|5.9.250.164|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]
Saving to: ‘index.html’
index.html [ <=> ] 13.76K --.-KB/s in 0s
2015-12-28 14:20:01 (421 MB/s) - ‘index.html’ saved [14093]
It seems to have downloaded it fine after one error, but that one error lasted the entire mini-outage, just sitting there. Meanwhile, ping, traceroute, and host were all fine. They didn't delay like wget did, and they didn't eventually give up and timeout like Firefox and the AUR helpers did. In fact, traceroute did even better than normal. Usually, it spends a while getting through those shitty Comcast servers, but this time, it just sped right through them instantly (even the 5000 ping one).
$ host aur.archlinux.org
aur.archlinux.org has address 5.9.250.164
aur.archlinux.org has IPv6 address 2a01:4f8:160:3033::2
aur.archlinux.org mail is handled by 10 mx.archlinux.org.
$ traceroute aur.archlinux.org
traceroute to aur.archlinux.org (5.9.250.164), 30 hops max, 60 byte packets
1 gateway (192.168.1.1) 5.494 ms 5.923 ms 5.919 ms
2 96.120.89.117 (96.120.89.117) 11.914 ms 14.856 ms 16.980 ms
3 te-0-7-0-3-sur03.scotts.ca.sfba.comcast.net (68.86.249.73) 18.317 ms 18.322 ms 18.319 ms
4 be-221-ar01.santaclara.ca.sfba.comcast.net (69.139.199.205) 19.507 ms 20.843 ms 23.648 ms
5 be-33651-cr01.sunnyvale.ca.ibone.comcast.net (68.86.90.93) 1955.554 ms 4615.323 ms 5025.058 ms
6 be-10925-cr01.9greatoaks.ca.ibone.comcast.net (68.86.87.158) 30.630 ms 21.116 ms 21.074 ms
7 he-0-14-0-1-pe03.11greatoaks.ca.ibone.comcast.net (68.86.86.202) 16.823 ms 13.608 ms 15.476 ms
8 ae12.sjc12.ip4.gtt.net (173.205.58.169) 22.706 ms 22.092 ms 21.607 ms
9 et-4-3-0.fra28.ip4.gtt.net (89.149.129.73) 176.661 ms 178.218 ms 177.570 ms
10 * hetzner-online-gw.ip4.gtt.net (77.67.76.142) 178.218 ms 178.184 ms
11 * core21.hetzner.de (213.239.245.217) 180.376 ms core22.hetzner.de (213.239.245.178) 179.523 ms
12 juniper1.rz16.hetzner.de (213.239.245.94) 171.577 ms 184.045 ms 184.261 ms
13 hos-tr3.ex3k3.rz16.hetzner.de (213.239.223.196) 182.020 ms hos-tr5.ex3k3.rz16.hetzner.de (213.239.251.74) 181.975 ms hos-tr1.ex3k3.rz16.hetzner.de (213.239.222.68) 173.405 ms
14 luna.archlinux.org (5.9.250.164) 180.403 ms 179.921 ms 180.375 ms
$ ping -nc3 aur.archlinux.org
PING aur.archlinux.org (5.9.250.164) 56(84) bytes of data.
64 bytes from 5.9.250.164: icmp_seq=1 ttl=41 time=174 ms
64 bytes from 5.9.250.164: icmp_seq=2 ttl=41 time=174 ms
64 bytes from 5.9.250.164: icmp_seq=3 ttl=41 time=292 ms
--- aur.archlinux.org ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2001ms
rtt min/avg/max/mdev = 174.406/213.787/292.042/55.336 ms
I ran all 4 commands at the same time as the AUR helpers in 4 different terminals (6 counting the helpers) to make sure it was all at the same time.
Last edited by Cadeyrn (2015-12-28 22:26:52)
Offline
I really confirmed that it's just the AUR that has the problem, and even archlinux.org works perfectly fine while I can't get to aur.archlinux.org
That's not particularly important. The front Arch page is hosted on a totally different server (actually, a totally different continent I believe) than the AUR.
I can't see any problems in the output you've posted so my theory has been shot down
My next guess is packet loss. Run this (while you're having problems) and post the output (it will take a while, with no progress output until it's done):
ping -qc100 aur.archlinux.org
Or even better, install `mtr` and run `mtr aur.archlinux.org` for a few minutes while you're having problems and post the output.
Are you familiar with our Forum Rules, and How To Ask Questions The Smart Way?
BlueHackers // fscanary // resticctl
Offline
I don't think we'll be able to get accurate results for the packet loss issue. I'm in Santa Cruz, in an area where Comcast is the only option, and they're infamous for being terrible. One problem that I already know I have is intermittent packet loss, much like the problem I'm having with the AUR, only with my entire internet connection. The only difference is, that issue only shows up a few times a day, whereas this AUR issue is sticking around constantly, all day, and only just showed up yesterday (as opposed to the Comcast issue that's always been around). I'd have to get lucky with my timing and somehow confirm exactly how lucky I got to know for sure whether the packet loss was from this or that. I'm willing to bet this is a packet loss issue, though.
I just ran the ping command you suggested and got 7% packet loss, but that aforementioned Comcast issue cropped up just as that ping was ending, so I don't know if that 7% came from that or from our AUR problem.
Last edited by Cadeyrn (2015-12-28 22:39:09)
Offline
This is still happening, and I'm convinced it's an issue with ipv6.
This will succeed every time:
> curl -4 "https://aur.archlinux.org/rpc?v=5&type= … arg=foobar"
This will timeout/hang almost every time:
> curl -6 "https://aur.archlinux.org/rpc?v=5&type= … arg=foobar"
Author of Aura
Offline
It's happened to me every now and then since, too. I've since figured out every internet problem I had that wasn't this AUR issue was in fact the fault of my using Wi-Fi, and switched to ethernet to find no problems whatsoever (except this one).
It's no longer often on my desktop, but it's still extremely often on my laptop, which is also running Arch Linux. It's so bad on my laptop that I've taken to updating its AUR packages by just copying them from my desktop's cache. But the weird part there is, my laptop gets internet through an ethernet cable to my desktop, and nothing in my local network uses IPv6. And yet there'll constantly be times where my laptop refuses to have a strong enough connection to use the AUR at all while at the same time it works fine on my desktop.
Last edited by Cadeyrn (2016-04-21 23:37:44)
Offline