The registration website crashed in time for the last day of "Early Bird Registration"
Too many people accessing the website at this time. Please try later.
Same here
I have tried all afternoon.
Dennis Chen Secrety Wikimedia Taiwan
On Sun, Apr 14, 2013 at 8:21 PM, Butch Bustria butch@wikimedia.org.phwrote:
The registration website crashed in time for the last day of "Early Bird Registration"
Too many people accessing the website at this time. Please try later.
-- Roman "Butch" Bustria Jr. Vice President (2012-2013)
Wikimedia Philippines Inc.
The information contained in this message is privileged and intended only for the recipients named. If the reader is not a representative of the intended recipient, any review, dissemination or copying of this message or the information it contains is prohibited. If you have received this message in error, please immediately notify the sender, and delete the original message and attachments.
Please consider the environment before printing this email.
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
Am 14.04.2013 14:47, schrieb Federico Leva (Nemo):
Butch Bustria, 14/04/2013 14:21:
The registration website crashed in time for the last day of "Early Bird Registration"
Yes, the centralnotice yesterday killed the site and was took down.
I am really disappointed by this. It should be obvious to everyone that a single server cannot cope with a centralnotice on Wikimedia projects, served by 300+ servers.
I have this discussions almost every month. Once WMIN is abusing the server, another time WLM, this time Wikimania. Each and every time nobody even thought a second to at least notify the sponsor of the server before doing something like this.
And I say this again and again every time, to no avail.
The problem can be avoided so easily - using a static HTML landing page (even PHP should work, just avoid any database transactions on the landing page) and talk before you do something. I was away for a few days (Jury meeting in London, weekend) and did not use a computer for two days, hence I could not see or react to anything. If I knew, I would've giving some hints beforehand and made sure I am monitoring the situation closely during that time.
/Manuel
Caching and Nginx are your first friends, every time.
Tom
On 14 April 2013 14:38, Manuel Schneider manuel.schneider@wikimedia.chwrote:
Am 14.04.2013 14:47, schrieb Federico Leva (Nemo):
Butch Bustria, 14/04/2013 14:21:
The registration website crashed in time for the last day of "Early Bird Registration"
Yes, the centralnotice yesterday killed the site and was took down.
I am really disappointed by this. It should be obvious to everyone that a single server cannot cope with a centralnotice on Wikimedia projects, served by 300+ servers.
I have this discussions almost every month. Once WMIN is abusing the server, another time WLM, this time Wikimania. Each and every time nobody even thought a second to at least notify the sponsor of the server before doing something like this.
And I say this again and again every time, to no avail.
The problem can be avoided so easily - using a static HTML landing page (even PHP should work, just avoid any database transactions on the landing page) and talk before you do something. I was away for a few days (Jury meeting in London, weekend) and did not use a computer for two days, hence I could not see or react to anything. If I knew, I would've giving some hints beforehand and made sure I am monitoring the situation closely during that time.
/Manuel
Wikimedia CH - Verein zur Förderung Freien Wissens Lausanne, +41 (21) 34066-22 - www.wikimedia.ch
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
Am 14.04.2013 15:40, schrieb Thomas Morton:
Caching and Nginx are your first friends, every time.
this is a general server for many different purposes. I don't have the money to sponsor a specialised setup just for two day of Wikimania registration and central notice.
/Manuel
Which begs the question; why have Wikimania HK not budgeted for custom IT??
(ps; if you're seeing these issues a lot it's still not a bad idea to chuck a cache up in front of the web server)
Tom
On 14 April 2013 14:45, Manuel Schneider manuel.schneider@wikimedia.chwrote:
Am 14.04.2013 15:40, schrieb Thomas Morton:
Caching and Nginx are your first friends, every time.
this is a general server for many different purposes. I don't have the money to sponsor a specialised setup just for two day of Wikimania registration and central notice.
/Manuel
Wikimedia CH - Verein zur Förderung Freien Wissens Lausanne, +41 (21) 34066-22 - www.wikimedia.ch
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
Am 14.04.2013 15:47, schrieb Thomas Morton:
Which begs the question; why have Wikimania HK not budgeted for custom IT??
(ps; if you're seeing these issues a lot it's still not a bad idea to chuck a cache up in front of the web server)
Sorry, this discussion is totally off-topic. We have done this for a while and it was never neccessary to have another extra IT just for Wikimania.
The question at hand is why people don't think about the consequences before they act, why central notices are created linking to directly to pages full of business logic without talking with the people who run that page first.
BTW, in regards to your last mail: That server already does file and memory caching but the website also has to be programmed appropriately to make the cache effective.
/Manuel
Oh, how they wish they had money to do that! Blame the WMF's being so stingy on the front of money in the last 2 years.
(personal opinion, not to be taken as Wikimania team's official position)
On 14 April 2013 14:47, Thomas Morton morton.thomas@googlemail.com wrote:
Which begs the question; why have Wikimania HK not budgeted for custom IT??
(ps; if you're seeing these issues a lot it's still not a bad idea to chuck a cache up in front of the web server)
Tom
On 14 April 2013 14:45, Manuel Schneider manuel.schneider@wikimedia.chwrote:
Am 14.04.2013 15:40, schrieb Thomas Morton:
Caching and Nginx are your first friends, every time.
this is a general server for many different purposes. I don't have the money to sponsor a specialised setup just for two day of Wikimania registration and central notice.
/Manuel
Wikimedia CH - Verein zur Förderung Freien Wissens Lausanne, +41 (21) 34066-22 - www.wikimedia.ch
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
On Mon, Apr 15, 2013 at 7:30 PM, Deryck Chan deryckchan@gmail.com wrote:
Oh, how they wish they had money to do that! Blame the WMF's being so stingy on the front of money in the last 2 years.
(personal opinion, not to be taken as Wikimania team's official position)
Why should the foundation have to spend money on a chapters IT setup?
On 15/04/2013 10:36, K. Peachey wrote:
On Mon, Apr 15, 2013 at 7:30 PM, Deryck Chan deryckchan@gmail.com wrote:
Oh, how they wish they had money to do that! Blame the WMF's being so stingy on the front of money in the last 2 years.
(personal opinion, not to be taken as Wikimania team's official position)
Why should the foundation have to spend money on a chapters IT setup?
Most of chapters' money already come from the foundation annual fundraiser either as FDC or grants program grants....
KTC
On Mon, Apr 15, 2013 at 7:42 PM, Katie Chan ktc@ktchan.info wrote:
Most of chapters' money already come from the foundation annual fundraiser either as FDC or grants program grants....
And they should budget that income and use appropriately, no? The previous comment makes it sounds they just want the foundation to keep handing money anytime someone wants it.
Am 15.04.2013 11:45, schrieb K. Peachey:
On Mon, Apr 15, 2013 at 7:42 PM, Katie Chan ktc@ktchan.info wrote:
Most of chapters' money already come from the foundation annual fundraiser either as FDC or grants program grants....
And they should budget that income and use appropriately, no? The previous comment makes it sounds they just want the foundation to keep handing money anytime someone wants it.
its not a matter of money anyways, the IT of Wikimania (at least scholarship / registration) is already paid by a Wikimedia chapter (Wikimedia CH), so it IS budgeted, paid for and everything. The issue here was just planning of ressources / proper use of the ressources. The setup was just never intented to bear the load of all people in all projects in any languages clicking on a Central Notice, that's all. If we knew before, we could have handled. In this case we could've even avoided the load without increasing IT ressources / budget.
/Manuel
Yes, at first I would like to thank the help from WM CH actually they given a crucial hand during the hour we figure out how to deal with the whole system thing
Also we are sure that WMF had given us pretty much needed support on having a successful event in HK although simultaneously, there is something quite unhappy happened during FDC or other grant stuff beyond the Wikimania budget
like saying we are not well organized, well tracked etc
but after all with other works with only volunteers with only 1/5 of time they can work for Wikimedia, how much you can expect, it is always a problem of Chicken and egg, and resources
On Mon, Apr 15, 2013 at 12:04 PM, Manuel Schneider < manuel.schneider@wikimedia.ch> wrote:
Am 15.04.2013 11:45, schrieb K. Peachey:
On Mon, Apr 15, 2013 at 7:42 PM, Katie Chan ktc@ktchan.info wrote:
Most of chapters' money already come from the foundation annual
fundraiser
either as FDC or grants program grants....
And they should budget that income and use appropriately, no? The previous comment makes it sounds they just want the foundation to keep handing money anytime someone wants it.
its not a matter of money anyways, the IT of Wikimania (at least scholarship / registration) is already paid by a Wikimedia chapter (Wikimedia CH), so it IS budgeted, paid for and everything. The issue here was just planning of ressources / proper use of the ressources. The setup was just never intented to bear the load of all people in all projects in any languages clicking on a Central Notice, that's all. If we knew before, we could have handled. In this case we could've even avoided the load without increasing IT ressources / budget.
/Manuel
Wikimedia CH - Verein zur Förderung Freien Wissens Lausanne, +41 (21) 34066-22 - www.wikimedia.ch
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
Thomas Morton, 14/04/2013 15:40:
Caching and Nginx are your first friends, every time.
Maybe the first, but in such cases usually the best is learning to link only a Wikimedia project page from the sitenotice or centralnotice. Notices get a lot of random clicks, few are interested in proceeding to the server where the meat is. Moreover, on our wikis we can use the Translate extension. I think this common sense rule may be added to https://meta.wikimedia.org/wiki/CentralNotice/Usage_guidelines. Only destinations tested for the load should be linked.
Nemo
Am 14.04.2013 16:12, schrieb Federico Leva (Nemo):
Maybe the first, but in such cases usually the best is learning to link only a Wikimedia project page from the sitenotice or centralnotice. Notices get a lot of random clicks, few are interested in proceeding to the server where the meat is. Moreover, on our wikis we can use the Translate extension. I think this common sense rule may be added to https://meta.wikimedia.org/wiki/CentralNotice/Usage_guidelines. Only destinations tested for the load should be linked.
Thanks Frederico.
Done right it is really no issue. WLM 2011 was not problem, neither Wikimania 2011 and WikiCon 2012.
But directly linking CMSes and MediaWiki sites from Central Notice are not a good idea. A quick notification would have helped to do it right. The lack of notification in conjunction with weekend and people being away from their computers is a bad combination.
/Manuel
This is going off topic (sorry) but I've been wondering for a while whether the chapters need to collaborate on a robust infrastructure for their technology needs - to increase capacity, reduce costs and deal with situations such as these (i.e. have expertise on hand and widely available).
At Wikimedia UK we're in the process of building a resilient infrastructure; with the modern capability of spinning up cheap servers behind load balancers there is no real need to have everything on a single piece of infrastructure. And the tools exist to scale horizontally if needed.
Tom
On 14 April 2013 15:15, Manuel Schneider manuel.schneider@wikimedia.chwrote:
Am 14.04.2013 16:12, schrieb Federico Leva (Nemo):
Maybe the first, but in such cases usually the best is learning to link only a Wikimedia project page from the sitenotice or centralnotice. Notices get a lot of random clicks, few are interested in proceeding to the server where the meat is. Moreover, on our wikis we can use the Translate extension. I think this common sense rule may be added to https://meta.wikimedia.org/wiki/CentralNotice/Usage_guidelines. Only destinations tested for the load should be linked.
Thanks Frederico.
Done right it is really no issue. WLM 2011 was not problem, neither Wikimania 2011 and WikiCon 2012.
But directly linking CMSes and MediaWiki sites from Central Notice are not a good idea. A quick notification would have helped to do it right. The lack of notification in conjunction with weekend and people being away from their computers is a bad combination.
/Manuel
Wikimedia CH - Verein zur Förderung Freien Wissens Lausanne, +41 (21) 34066-22 - www.wikimedia.ch
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
Sure it exists, it is my daily work and bread and butter. I just feel like I am already sponsoring enough free hosting for a lot of chapters. Fortunately WMCH and WMAT picked up now and started hosting their own servers with me - but still, these are single servers used for many different purposes. But cheap, sure (100 EUR / month on a cloud infrastructure).
/Manuel
Am 14.04.2013 16:57, schrieb Thomas Morton:
This is going off topic (sorry) but I've been wondering for a while whether the chapters need to collaborate on a robust infrastructure for their technology needs - to increase capacity, reduce costs and deal with situations such as these (i.e. have expertise on hand and widely available).
At Wikimedia UK we're in the process of building a resilient infrastructure; with the modern capability of spinning up cheap servers behind load balancers there is no real need to have everything on a single piece of infrastructure. And the tools exist to scale horizontally if needed.
Just to return to the original subject, my application was accepted (I have the email confirming it), but for some reason my paypal account was not debited. I shall leave it to the Wikimania organizers to resolve this (I've written to them directly), but it might be an idea for others to check their paypal accounts to ensure that they were properly debited.
Risker/Anne
You should have got 2 e-mail for confirmation, One for registration received, another for registration confirmed (payment received).
Justin
Just to return to the original subject, my application was accepted (I have the email confirming it), but for some reason my paypal account was not debited. I shall leave it to the Wikimania organizers to resolve this (I've written to them directly), but it might be an idea for others to check their paypal accounts to ensure that they were properly debited.
Risker/Anne
Wikimania-l mailing list Wikimania-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimania-l
wikimania-l@lists.wikimedia.org