EVE Technology Lab

Thanks Vaerah. I was wondering if it was just me that cannot access the site :). Today I'm constantly getting connection timeouts.

Balls, it seems the site either crashed, or someone at my datacenter did something incredibly stupid, since my server died (presumably) without reason some 14 hours ago. I'm currently investigating this issue.

UPDATE: Oh joy, the server is moving, and my hosting company does not know why...

---

Your links are broken;

* dig chubaka.dyndns.biz +short
35.32.132.87
* ping -c 1 35.32.132.87
PING 35.32.132.87 (35.32.132.87) 56(84) bytes of data.
--- 35.32.132.87 ping statistics ---
1 packets transmitted, 0 received, 100% packet loss, time 10000ms

Sounds like you need to fixy your dynamic dns service? :)

I'm sure it's all related to your current posts, but the following address: http://www.evemarketeer.com/api/averages/10000002/30/620/xml

Gave the following error: "Unable to connect to the remote server

Not sure if that helps

Thanks for looking into it.

Do you all fail on reading that his server is down?

I can read.

dig www.evemarketeer.com ns
;; Got bad packet: extra input data
101 bytes
2d 0f 81 80 00 01 00 00 00 01 00 00 03 77 77 77 -............www
0c 65 76 65 6d 61 72 6b 65 74 65 65 72 03 63 6f .evemarketeer.co
6d 00 00 02 00 01 c0 10 00 06 00 01 00 00 38 40 m.............8@
00 33 00 00 0a 68 6f 73 74 6d 61 73 74 65 72 0c .3...hostmaster.
65 76 65 6d 61 72 6b 65 74 65 65 72 03 63 6f 6d evemarketeer.com
00 77 df 4d 38 00 00 38 40 00 00 0e 10 00 12 75 .w.M8..8@......u
00 00 00 38 40 ...8@

^ Get a better DNS resolver?

---

dig www.evemarketeer.com ns @ns1.coronum.nl +short

; <<>> DiG 9.7.3 <<>> www.evemarketeer.com ns @ns1.coronum.nl +short
;; global options: +cmd
;; connection timed out; no servers could be reached

---

dig evemarketeer.com a +short
89.31.102.172 (redirects)
dig www.evemarketeer.com a +short
50.28.89.81 (site loads)

Why not change the @ DNS to point to the new ip address? I think you should get a better name server resolution, there are several out there. I'd suggest; http://zoneedit.com

Or failing that.. http://freedns.afraid.org/ :)

Edit: Looks like your on liquid web;
http://www.liquidweb.com/dedicated/single-processor.html

Why not switch to a more reliable VPS provider like linode? Just curious.

NGINX is a lightweight http daemon service.

Bad Gateway seems to mean that how it's connected with the PHP parser is not syncing properly, there are many reasons this happens. I don't know your config, are you using nginx for everything? A reverse proxy with apache? or something else?

I still think you could get a better host. The whole reason people use nginx in the first place is because they lack enough ram to use Apache reliably.

Prices are now about 7 days old according to the little green/yellow/orange dots.

http://www.evemarketeer.com/item/group_summary/18

There's also close to 100k in queued uploads.

Because Liquidweb is better, incredibly less expensive and faster. Those are just matters of opinion, but they count for a lot with me.

The whole reason people use Nginx, is that there is little reason to load PHP into memory just to serve an image. When using apache, there is little reason not to use mod_php, so...

In general, I just feel nginx is the better of the two for my purposes. No, wait, that is not it. More like... I just felt like it at the time.

Though to be fair, you were right about something being wrong with DNS.



As you might have noted, the server is now up again, though it still seems to not be processing any uploads (or not saving or showing it anyway). I'm going to investigate that now.

I'm sorry for my inability to really take care of the website, but there are just more important things in my life right now.

UPDATE: The problem that was causing the non-processed uploads was a 30GB task queue, caused by my inadvertently not turning on a certain processing task (the one dealing with the information you see on pretty much any page), every upload caused about 6 or 7 of these to spawn in the task queue, so it quickly became clogged up, and mysql just couldn't deal with the enormous table anymore.

I will also be switching over to a SSD based server sometime soon, hopefully making all the disk access mysql must do an order of magnitude faster.

UPDATE 2: The server is now smart enough to restart processing itself if it by any chance decides to shutdown and restart for any reason. I hope this will protect us against the humongous amounts of processing that had to be done after such occurrences.

UPDATE 3: Hopefully, the 502 warnings are now also gone for the forseeable future.

UPDATE 4: If anyone ever wanted to work on EVE Marketeer but thought it wasn't possible because he couldn't get the data, here you go: http://www.evemarketeer.com/var/dump/

It contains an export of all non sensitive data that is required to make the site work. If you always wanted to work with huge sets of data, this is for you as well :P

Another noobie question. For the first ime I have looked at the market orders. They website says they were updated 1 minute before I read them.

Yet in game I see Nocx @Jita CNAP at 985 or so on sell orders, but I see a number of website sell orders @Jita CNAP on the website for about 531 ISK pu before the 985 ones. What gives?