| Pages: 1 [2] :: one page |
| Author |
Thread Statistics | Show CCP posts - 4 post(s) |
Lindsay Logan
   |
Posted - 2008.09.27 14:36:00 -
[31]
Edited by: Lindsay Logan on 27/09/2008 14:36:32
omg! What is up with ppl!?
CCP have a technical issue, its not the end of the world. They did not intend this, and is working to fix it. Is that so hard to understand!?
And for information. I got on ok, no problem, but that was 8 min after DT.
|
Cevin North
|
Posted - 2008.09.27 14:37:00 -
[32]
it works?!?! Server showed offline for 30 seconds on evemon and then it started working again
|
Angry Cyno
|
Posted - 2008.09.27 14:38:00 -
[33]
To shabby hardware to handle the players.. ccp you FAIL!
|
Maria Kalista
|
Posted - 2008.09.27 14:47:00 -
[34]
 Originally by: Angry Cyno
To shabby hardware to handle the players.. ccp you FAIL!
No one stops you from doing it better.
Computers and programs running on them do tend to break in the most inconvenient ways and moments.
Originally by: CCP Mitnal
You put a bear in your tea???
|
Ra Vhim
Black Bag Ops
|
Posted - 2008.09.27 14:51:00 -
[35]
OMG. I lost like an hour of training time due to this. The horror! This is an end of the world kind of thing. Must I kill myself or someone at CCP because such an outrageous thing can happen?
This is the right thread to voice my concerns, right? I don’t know if I can live with the knowledge that someday, without any prior notice, I might have trouble connecting to Eve. The mental stress may drive me mad. If I have a mental breakdown I will sue CCP, I promise. It isn’t healthy to constantly bear the burden of the knowledge that I suddenly, without warning, might be mentally assaulted and violated by a login queue!
P.S.
The troubles of the western world rocks. It is like living in a soap opera.
|
|
CCP Navigator
C C P
|
Posted - 2008.09.27 15:12:00 -
[36]
Hello everyone,
Our Operations Team have resolved the issue and I wanted to provide you with an update as soon as possible.
Our load balancing hardware that feeds connections to our proxy servers was feeding too many connections (7000) to one machine where as our other proxies where below 1000. This caused the players that where added to that proxy to enter a login queue. We have resolved the matter by shutting down that proxy and are analysing what configuration in our load balancing hardware is causing this abnormal behaviour.
Navigator
Senior Community Representative
CCP Hf, EVE Online
Email / Netfang
|
|
Angry Cyno
|
Posted - 2008.09.27 15:42:00 -
[37]
MSHD-4?
Nodecrash = you fail again CCP
|
Misanth
RABBLE RABBLE RABBLE
|
Posted - 2008.09.27 15:43:00 -
[38]
Respect for keeping us updated fast, as well as finding the issue fairly quick at a weekend, not bad.
|
DoctorSeuss
|
Posted - 2008.09.27 15:55:00 -
[39]
I pay to stare at a black screen, don't you all?
|
malet
Quam Singulari
|
Posted - 2008.09.27 15:56:00 -
[40]
thanks CCP
All fine now :)
QUAM - Hot dropping a system near you soon
|
DigitalCommunist
Obsidian Core
|
Posted - 2008.09.27 16:11:00 -
[41]
Originally by: SharpMango
Just out of curiousity.
if i get podded in space because i got a random DC upon connecting..
and foolishly (oh so foolishly undocking  )...
are the game mechanics working as intended?
I AM NOT RANDOM
|
Haywoode Jablome
State War Academy
|
Posted - 2008.09.27 16:46:00 -
[42]
Originally by: CCP Navigator
Hello everyone,
Our Operations Team have resolved the issue and I wanted to provide you with an update as soon as possible.
Our load balancing hardware that feeds connections to our proxy servers was feeding too many connections (7000) to one machine where as our other proxies were below 1000. This caused the players that where added to that proxy to enter a login queue. We have resolved the matter by shutting down that proxy and are analysing what configuration in our load balancing hardware is causing this abnormal behaviour.
Hmmm, your servers suck... it should be OVER 9000!
|
Verx Interis
Amarr
Aurora Security
Cosmic Anomalies
|
Posted - 2008.09.27 16:49:00 -
[43]
Well, it's fixed now. And the amount of people who said they'll ragequit because they lost a few hours to a day of skills is just as entertaining.
---------
Friggin signature size limit.
The Eve professions list |
BiggestT
Caldari
Space Oddysey
Pupule 'Ohana
|
Posted - 2008.09.27 17:43:00 -
[44]
Originally by: Verx Interis
Well, it's fixed now. And the amount of people who said they'll ragequit because they lost a few hours to a day of skills is just as entertaining.
fixed is it?
Is that why im stil lstaring at a black screen?
Awesome EVE history
|
Lindsay Logan
|
Posted - 2008.09.27 17:44:00 -
[45]
Originally by: Angry Cyno
MSHD-4?
Nodecrash = you fail again CCP
I don't see you running very successful MMO for 5 years. Go back to your corner troll.
|
BiggestT
Caldari
Space Oddysey
Pupule 'Ohana
|
Posted - 2008.09.27 17:57:00 -
[46]
Edited by: BiggestT on 27/09/2008 18:01:46
In, ty :D
Awesome EVE history
|
James Forsyth
Gallente
|
Posted - 2008.09.27 18:28:00 -
[47]
Thank you for quickly addressing this issue.
Veritas vos liberabit. |
MotherMoon
Huang Yinglong
|
Posted - 2008.09.27 22:47:00 -
[48]
Originally by: Angry Cyno
MSHD-4?
Nodecrash = you fail again CCP
wrangler allready said they moved the jita node to you now becuase of the crashes, be happy.
|
SharpMango
14th Legion
|
Posted - 2008.09.27 23:04:00 -
[49]
Originally by: CCP Navigator
Originally by: SharpMango
Just out of curiousity.
if i get podded in space because i got a random DC upon connecting..
and foolishly (oh so foolishly undocking )...
are the game mechanics working as intended?
That is an issue that you would need to raise via the petition system. Only a GM can answer that query fr you if it were to happen.
thx for the response,
im chuffed it didnt happen though, they dont make pirates like they used to lol
|
NereSky
Gallente
Trinity Nova
Trinity Nova Alliance
|
Posted - 2008.09.28 11:44:00 -
[50]
Originally by: Audrea
Originally by: CCP Navigator
Originally by: SharpMango
Just out of curiousity.
if i get podded in space because i got a random DC upon connecting..
and foolishly (oh so foolishly undocking )...
are the game mechanics working as intended?
That is an issue that you would need to raise via the petition system. Only a GM can answer that query fr you if it were to happen.
Right, Our logs, they show nothing!
yeah or they will start the reply with 'sorry to hear about the loss of your ship.........' (even though it was implants you may have lost lmao
|
Xiaodown
coracao ardente
Triumvirate.
|
Posted - 2008.09.29 13:35:00 -
[51]
CCP:
Round Robin DNS + KeepaliveD + iproute2 source routing = total win.
Customer -> multiple DNS entries -> Multiple routers running keepalived -> Multiple proxies.
Round Robin DNS will allow people to hit any one of a number of routers, which can be synched with VRRP to have IPs float between them for the event of a failure.
Running IPVS / KeepaliveD on the firewall will allow the "Virtual IP" (i.e. the publicly addressable IP) to be mapped to any number of "real servers", so you can map 1.2.3.4 -> 192.168.100.X where X is 1,2,3,4,5,6, however many proxies you have. Keepalived will also perform "health checks" on the back end servers and automatically pull them out of rotation under certain circumstances.
Part of the problem is that with multiple routers forwarding to multiple back-end proxies (or back-end servers in general), you will run into TCP loop conditions; by that I mean, if you come in via Firewall3, but your proxy is set with a default gateway of Firewall1, your customers will never get their TCP ACKs back. This can be remedied by running one IP per proxy per firewall. For example, for 3 proxies, you can have 192.168.101.(1,2,3) be for proxy(1,2,3), for firewall1, and 192.168.102.(1,2,3) be for proxy(1,2,3) for firewall2, etc. Then, with iproute2 you can set up entries in /etc/rc_tables for each firewall, and set up custom iproute rules on your new IPs:
/sbin/ip rule add from 192.168.101.0/24 table firewall1
/sbin/ip rule add from 192.168.102.0/24 table firewall2
(etc...)
/sbin/ip route add default via <firewall1's ip> dev eth0 table firewall1
/sbin/ip route add default via <firewall2's ip> dev eth0 table firewall2
Which will make the routing table look up the corresponding rule for traffic coming into any one of the new IPs. That way, traffic that lands on 192.168.103.2 will be traffic coming from firewall3 to proxy2, and will be routed back out firewall3 (for traffic matching that CIDR range, it will look up table firewall1, which will tell it that it's default interface is via firewall3's IP).
This is contingent on the proxy software being able to bind to all IPs on the system.
Bickity bam, high availability and load balancing.
/Rackspace Linux High Availability sysadmin
/Coming to London sometime soon, willing to consult!
|
| |
|
| Pages: 1 [2] :: one page |
| First page | Previous page | Next page | Last page |