imgur.com been slow all week now unreachable

[#1]

Don't know if this helps but my wireshark logs
If anyone is having a problem with imgur; https everywhere plugin for firefox or just using https:/www.imgur.com has fixed it for me,
I think imgur is filtered by the IWF but as they hve millions of users and are trying to fileter specific urls the plusnet or iwf cache is failing, would be good to know if plusnet or iwf are causing the problem.


GET / HTTP/1.1
Host: www.imgur.com
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:35.0) Gecko/20100101 Firefox/35.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US,en;q=0.5
Accept-Encoding: gzip, deflate
Connection: keep-alive

HTTP/1.0 302 Moved Temporarily
Retry-After: 0
Location: http://imgur.com/
Content-Length: 0
Accept-Ranges: bytes
Date: Sat, 30 Jan 2016 14:33:51 GMT
X-Served-By: cache-lhr6320-LHR
X-Cache: HIT
X-Cache-Hits: 0
Server: cat factory 1.0
Cache-Control: no-store, no-cache, must-revalidate, post-check=0, pre-check=0
X-Frame-Options: DENY
X-Cache: MISS from pcl-iwfcache01.plus.net
X-Cache-Lookup: MISS from pcl-iwfcache01.plus.net:80
Via: 1.0 pcl-iwfcache01.plus.net:80 (squid)
Connection: close


Source: pcl-iwfcache02.plus.net (212.159.3.234)
83 55.866258000 pcl-iwfcache02.plus.net 192.168.1.46 ICMP 90 Time-to-live exceeded (Time to live exceeded in transit)

GET / HTTP/1.1
Host: www.imgur.com
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:35.0) Gecko/20100101 Firefox/35.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US,en;q=0.5
Accept-Encoding: gzip, deflate
Cookie: SESSIONDATA=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Connection: keep-alive
HTTP/1.1 302 Moved Temporarily
Retry-After: 0
Location: https://imgur.com/
Content-Length: 0
Accept-Ranges: bytes
Date: Sat, 30 Jan 2016 16:25:55 GMT
Connection: close
X-Served-By: cache-lhr6335-LHR
X-Cache: HIT
X-Cache-Hits: 0
Server: cat factory 1.0
Cache-Control: no-store, no-cache, must-revalidate, post-check=0, pre-check=0
X-Frame-Options: DENY
21 7.146760 192.168.1.46 185.31.19.193 TCP 54 43939 → 443 [ACK] Seq=1 Ack=1 Win=66560 Len=0
GET / HTTP/1.1
Host: imgur.com
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:35.0) Gecko/20100101 Firefox/35.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US,en;q=0.5
Accept-Encoding: gzip, deflate
Cookie: SESSIONDATA=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Connection: keep-alive
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
Cache-Control: max-age=60, stale-while-revalidate=600, stale-if-error=86400, public
Fastly-Debug-Digest: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Content-Length: 151913
Accept-Ranges: bytes
Date: Sat, 30 Jan 2016 16:25:56 GMT
Age: 109
Connection: keep-alive
X-Served-By: cache-iad2127-IAD, cache-lhr6324-LHR
X-Cache: HIT, HIT
X-Cache-Hits: 2, 3
Vary: Accept-Encoding
Server: cat factory 1.0
X-Frame-Options: DENY

[#2]

Yeah i have been using the https everywhere plugin for chrome and it works fine, easy to disable also if not needed/wanted.. Sadly no use if i want to surf in bed from my ipad or iphone though!

[#3]

Quote from: Ravencroft

Can anyone suggest a free vpn I can use till this fixes? I pay a subscription to nexusmods.com and of course I can't even use it at the moment.

I downloaded CyberGhost since it was listed as free and doesn't throttle the connection, free version worked fine for me to be able to download 19gb worth of client data for TSW.  Free version has a 3 hour limit for the connection then you need to reconnect (and there's normally a queue that starts around 3000ish to get a connection but the queue number went down to 0 in < 2 mins and then I was connected. http://www.cyberghostvpn.com/en_gb?utm_expid=85864364-20.eNBY82fVRQSTe-9mpnQtJw.0

[#4]

Oh hey after the national lottery site problem was fixed another site breaks
Can't be anything but a plusnet issue if https works as fast as it can and http doesnt...
Meh, the time it takes between issues cropping up and getting fixed with PN is pretty poor 😐

[#5]

Thanks for the temp fix Jpixton, looks like we also need to add i.imgur.com to the list of HTTPS rules, as Reddit often links to this along with imgur.com

[#6]

There are also a bunch of folks having the same problem with www.nexusmods.com .  I posted a question on another thread here and have kindly been referred to here.

[#7]

Has anyone tried changing their DNS? That may mean that you go to a different cdn.

C:\Users\John>nslookup i.imgur.com
Server:  cdns01.plus.net
Address:  212.159.6.9
Non-authoritative answer:
Name:    imgur.com
Address:  185.31.18.193
Aliases:  i.imgur.com

C:\Users\John>nslookup i.imgur.com 8.8.8.8
Server:  google-public-dns-a.google.com
Address:  8.8.8.8
Non-authoritative answer:
Name:    imgur.com
Address:  185.31.19.193
Aliases:  i.imgur.com

[#8]

It isn't a DNS issue, I use google DNS as standard on all my devices and router

[#9]

I wasn't saying it was a DNS issue. If the DNS lookup from a different server gives the IP for a different cdn server the route to that server may be different and so may not be affected

[#10]

Can confirm I've had the same issue using both opendns and plus nets own dns servers so sadly not a solution

[#11]

Quote from: ejs

This was not a DNS issue. The IP address of imgur.com has not changed. There was no problem resolving the IP address. OK now I've noticed it's got two IP addresses, 185.31.18.193 and 185.31.19.193. You will get one or other of those IP addresses regardless of if you use Plusnet's DNS servers or Google DNS.

You get 185.31.18.193 or 185.31.19.193 at random regardless of what DNS servers you use. For imgur.com specifically, another workaround might be to add an entry to your hosts file with whichever of those IP addresses doesn't go through the IWF filter, which was apparently the one with .19.

[#12]

Strangely I see it the other way round:
traceroute to 185.31.19.193 (185.31.19.193), 64 hops max, 52 byte packets
1  192.168.29.1 (192.168.29.1)  2.470 ms  1.143 ms  1.966 ms
2  195.166.130.220 (195.166.130.220)  8.894 ms  8.358 ms  8.925 ms
3  link1-central10.pcl-ag07.plus.net (84.93.249.193)  10.209 ms  11.845 ms
    link2-central10.pcl-gw02.plus.net (84.93.249.194)  8.378 ms
4  ae2.ptw-cr01.plus.net (195.166.129.4)  42.773 ms  9.826 ms
    ae1.pcl-cr02.plus.net (195.166.129.3)  13.295 ms
5  tun0.ptn-gw01.plus.net (212.159.2.101)  11.254 ms
    te9-4.ptn-gw01.plus.net (195.166.129.33)  12.633 ms
    tun0.ptn-gw01.plus.net (212.159.2.101)  9.266 ms
6  po2.pcl-gw02.plus.net (195.166.129.43)  9.983 ms
    tun0.ptn-gw01.plus.net (212.159.2.101)  9.908 ms
    po2.pcl-gw02.plus.net (195.166.129.43)  12.102 ms
7  tun0.pcl-gw01.plus.net (212.159.2.105)  10.484 ms
    flowers2.servers.plus.net.uk (212.159.2.98)  9.524 ms
    tun0.pcl-gw01.plus.net (212.159.2.105)  8.481 ms
8  195.99.126.138 (195.99.126.138)  9.482 ms
    flowers2.servers.plus.net.uk (212.159.2.98)  74.771 ms
    195.99.126.138 (195.99.126.138)  9.101 ms
9  195.99.126.96 (195.99.126.96)  10.655 ms
    core3-te0-0-0-14.faraday.ukcore.bt.net (213.121.193.160)  9.967 ms
.......
as opposed to
traceroute to 185.31.18.193 (185.31.18.193), 64 hops max, 52 byte packets
1  192.168.29.1 (192.168.29.1)  1.865 ms  1.167 ms  1.473 ms
2  195.166.130.220 (195.166.130.220)  8.184 ms  7.731 ms  8.005 ms
3  link1-central10.pcl-ag07.plus.net (84.93.249.193)  8.102 ms  10.398 ms
    link2-central10.pcl-gw02.plus.net (84.93.249.194)  8.776 ms
4  195.99.126.138 (195.99.126.138)  10.135 ms  11.538 ms
    195.99.126.182 (195.99.126.182)  11.368 ms
5  core1-te0-3-0-12.ealing.ukcore.bt.net (109.159.254.101)  12.047 ms
    host213-121-193-153.ukcore.bt.net (213.121.193.153)  11.876 ms  12.549 ms
6  core2-te0-11-5-0.ealing.ukcore.bt.net (213.121.193.74)  13.255 ms
    core1-te0-4-0-5.ealing.ukcore.bt.net (62.172.103.189)  11.596 ms
    host213-121-193-230.ukcore.bt.net (213.121.193.230)  11.487 ms
......
I have just put some entries into my routers /etc/hosts (yay for linux based routers and dnsmasq):
185.31.18.193    i.imgur.com
185.31.18.193    m.imgur.com
185.31.18.193    imgur.com
And seems to be working better for now, until for some reason traffic to that IP starts going through the proxy as well.....or imgur change IPs and I spend a few days shouting at things until I remember those addresses are hard coded in my router 

[#13]

FWIW, this is affecting PN and whitelabel customers (e.g. John Lewis Broadband). Was trying to figure out why it was so slow on parent's connection earlier, then I got back home and had exactly the same thing!

[#14]

Testing www.cyberghostvpn.com/ with affected websites and games and up shot is two fold.  It works but shows me that plusnet are capping download speeds.   e.g. warframe during its patching I get at best 1MB/s which also results in corrupted patches, I get the normal 3MB/s if using a UK vpn with Cyberghostvpn and no corrupting.
So bob or who ever what is going on, are you becoming Virgin Media and adding speed caps as you don't have the bandwidth.

[#15]

Flagging this up internally (assuming someone hasn't already), will update when we get more info.