Rosetta Code:Village Pump/Income: Difference between revisions
→Objections and other concerns: More information
(→Dear Wikimedia Foundation: Direct competitor) |
(→Objections and other concerns: More information) |
||
Line 50:
I assume this won't be for profit, but how much do you think you need to make to cover the costs? --[[User:Mwn3d|Mwn3d]] 15:21, 7 July 2009 (UTC)
: Right now, in static costs, Rosetta Code costs $25/mo. That only covers what I pay to Slicehost. The domain is paid for for the next ten years, so there's no extra cost there. I would ''like'' to get an additional slice, either as a doubling of the existing slice's capabilities, or as a separate server to provide a degree of redundancy or service separation. (It would be nice, for example, for me to be able to put Squid and memcached on one vps box while having php and mysql on another, or even to have a third vps slice dedicated to MySQL. But each vps slice adds roughly $20/mo to the cost.) Rosetta Code is not running me into a pit; I've got other costs I can cut if I discover I'm habitually in the red. But it's not something I can afford to dedicate more to, until either I make something more than $8/hr, or it starts paying for itself. I seriously dislike the idea of panhandling, which is why I'm looking for ways of getting Rosetta Code to pay for itself. My top idea right now is to sell user-defined books through a POD publisher, but I haven't found anyone who allows for programmatic submission of materials for print. --[[User:Short Circuit|Short Circuit]] 21:00, 7 July 2009 (UTC)
: Someone asked what kind of traffic Rosetta Code was dealing with, and why $25/mo wasn't handling it very well. In the interest of avoiding repeating it, I'll repeat it here:
:''At the moment, the load average is 0.15 0.04 0.01. That raises to between 1 and 2 while ImplSearchBot is running. Gets higher if someone triggers something that I could have configured to avoid. (For example, my load average shot through the roof when someone was repeatedly trying to view the thumbnail of the Dragon Curve animated GIF; ImageMagick took too long to render the thumbnail for caching, and they kept hitting Refresh.)''
:''My bigger problem is when I get traffic spikes. Every now and then, some major tech blog or news website will link to us. It's nowhere near Slashdot levels, but that much traffic exhausts 256MB of RAM pretty quickly. I don't like reducing PHP's internal memory limit, as that causes errors when large pages are submitted or rendered. (Which is the one technical reason we can't have an in-wiki table that shows a direct cross-section of all the languages with all the tasks; It's just too much data.)''
:''Normal load, it's fine, almost dead. When ImplSearchBot runs, it slows down a bit, but that only takes about ten minutes. But two or three times a year, I get enough outside traffic, or a particularly nasty query, that causes issues.''
:''And there are a few features that I want to add that I'm hesitant to because of the impact they'll have on server load. Things like extended-function wiki themes, a possible XUL interface, and even possibly a console-tuned interface that uses the MediaWiki API.''
:''Anyway, the vast majority of my problems don't come from not having enough disk, bandwidth or CPU, they come from not having enough RAM.'' --[[User:Short Circuit|Short Circuit]] 00:26, 8 July 2009 (UTC)
| |||