New record, new server, new hope, new costs
The site is hosted by Linode now.
Update:New goal – to upgrade the VPS to 1.5GB RAM (currently 1GB), 60GB (currently 40GB) disk space and to get a 10% discount we need $670 yearly.
Thanks much again to those who already donated!
If you find the site interesting and, hopefully, useful, I would greatly appreciate if you could consider making a donation to keep it running. Our goal is $670 (annually). Thank you for your help!
On Jan 18 we had achieved a new record: 62 on-line users in the forum. But the shared hosting could not deal with such a load and the site was blocked. After several sleepless nights the site has been finally moved to a new hosting plan, which is 10 times more expensive than the previous one $480 per year or $50 per month (DNS expenses not counted).
But now we have a virtual server with 1GB memory, 2 core 2.4GHz CPU, 40GB disk space and unlimited traffic. Not everything is working smoothly, but I’m doing my best on getting the issues resolved.
Please let me know if you notice something not working or if you have ideas.
Related articles and threads:
Related posts:
- The VPS has been migrated to a new hosting company!
Updated on Dec 04 @ 18:00The VPS has been migrated to a new hosting company – Linode! New OS, new... - Aaaand another downtime…. Or sysadmin self-bondage
After a 14hour long battle with the hoster the problem has been sold. The VPS was inadvertently moved to another... - 4 hour downtime
The physical server on which the VPS likera.com is located had some technical problems. As a result the disk system... - Maintenance and downtime
Following and upgrade and reboot the server was down for 6 hours. But finally the kernel has been fixed (thanks... - The recent 1 hour outage was caused by a ripper. Demotivator of the day
One idiot was ripping the site with enormous speed and connections per second. Something I haven’t tuned the site for,...
24 Responses to “New record, new server, new hope, new costs”
Blog updates
Likera, why don’t you use hazenet?
Like 8 bucks a month for a quad core 1gb ram.
Comment :: January 23, 2011 at 02:51 :: Quote
I’ve read very bad reviews about HazeNET. Arvixe (I’m currently using) scored one of the best (if not the best). They offer excellent support, response time and uptime. The prices are quite high, though.
But that does not mean that I’ve stopped looking around ;-)
Comment :: January 23, 2011 at 03:23 :: Quote
[...] if you have money to buy these style objects you definitely have some spare money to donate to this site ;-) #gallery-1 { margin: auto; } #gallery-1 .gallery-item { float: left; margin-top: 10px; [...]
Pingback :: February 11, 2011 at 02:21 :: Quote
Paid $50 ($20 from the pool + $30 refund) for the next 30 days.
Comment :: February 16, 2011 at 23:54 :: Quote
-$47 for March.
Comment :: March 18, 2011 at 00:29 :: Quote
My own, small, contribution on its way!
MJ
Comment :: April 20, 2011 at 14:53 :: Quote
Got it! Updated! Thanks much MadJack!
Comment :: April 20, 2011 at 15:02 :: Quote
Donated 25€ two days ago, thanks for running this site!
Comment :: April 21, 2011 at 20:01 :: Quote
Thank you guys again for your support!
Comment :: April 22, 2011 at 00:19 :: Quote
Here’s an idea, have users donate spare processing power while visiting your site to generate bitcoins for you! Use http://bitjam.org/ and probably just make a wallet at http://www.mybitcoin.com. You should probably look up bitcoins first and figure out how they work…
Comment :: June 14, 2011 at 03:21 :: Quote
Sure, will have a look. But do you know if it’s considered as a spy or a malware by search engines, web-proxies and antivirus software? Do not want to be blocked by half of the Internet. Took me years to shake off one ad-campaign residuals…
Comment :: June 14, 2011 at 12:42 :: Quote
I have it on my site and it’s clean… You might want to put an announcement if you do use it, so then people can “donate” by just leaving their computers on the web page… Some people do that for us, they just leave it on during work or at night.
Comment :: June 15, 2011 at 02:35 :: Quote
Alright, better idea… http://kradminer.com/
To have it automatically put your address in, use this link http://kradminer.com/?mine_for=13buRFcMxDtxFZHeEAFJ6K92aXoagxrBVP (of course replacing that one with yours).
Comment :: June 19, 2011 at 18:27 :: Quote
Maybe you could put up Flattr.com-buttons on the website too? I could sure click a dollar or two if needed. :)
Comment :: July 2, 2011 at 00:38 :: Quote
I’ve read about Flattr and was not impressed, to be honest. But I can try it, sure. Thanks for the suggestion!
Comment :: July 3, 2011 at 22:37 :: Quote
Added the Flattr button.
Comment :: July 24, 2011 at 03:33 :: Quote
Hey I don’t know if this will help but its related to moneyz.
Maybe you can give self-bondage advice for 5 bucks each.
I know fiverr.com doesn’t allow adult but this is just an idea. My 2 cents.
Here is a short article similar to my idea.
http://marketerswap.com/forum/showthread.php/464-Fiverr-and-Using-Other-Sites-Like-it-to-Earn-Coin?p=1076#post1076
Comment :: August 10, 2011 at 03:42 :: Quote
“Cal ” wrote:
This contradicts the idea behind the site.
BTW, the content is not showing up in Firefox.
Comment :: August 20, 2011 at 03:33 :: Quote
I’m always surprised by the resource requirements of this blog. I have 50,000-daily-active-user sites doing much more resource intensive work that use less. I have old phpBB forums with more posts and members than this one sharing resources with a dozen other sites on a 6 year old server. Perhaps if you switched from what’s probably Apache and mod_php to nginx and php-fpm…
At least I’m glad you moved to Linode. They’re a good choice for VPS. I use them for development and staging servers.
Comment :: May 9, 2012 at 02:39 :: Quote
“Dan ” wrote:
The problem is the disk which is shared by all VPSes. If the disk is abused by, for example, a backup, system installation or extensive swapping, all related VPSes suffer.
See also this thread: http://www.likera.com/forum/mybb/showthread.php?tid=530
Comment :: May 9, 2012 at 15:27 :: Quote
Disk IO should never be a bottleneck for a small blog or forum… it’s not like you’re uploading gigabytes of new files every minute. Most requests should be served straight from memory with the MySQL query cache, OS disk cache, and PHP opcode cache (APC) keeping what’s needed there automatically.
Comment :: May 9, 2012 at 21:22 :: Quote
“Dan ” wrote:
Sure, if it’s a standalone system. The site can survive a much higher load, if the disk load is normal. And at Arvixe the disk load was 100% even when the web-server was stopped. The whole story is in that forum thread (see above).
“Dan ” wrote:
95% of the site is dynamic. With quite a few write operations (view counters, for example). Some side widgets are fully cached, some (e.g. 4 random images) are not.
“Dan ” wrote:
Yes, true. The DB is not that big (130 tables, 65MB for this site). But caching takes memory, right?
“Dan ” wrote:
Yes, but cache takes memory.
“Dan ” wrote:
It’s set up to 64MB. The hit rate is quite high, but spiders and rippers kill the caches.
The available memory is only 1GB. If the disk is busy, apache processes accumulate in memory, what leads to swapping, swapping adds to the disk load, etc. I have a script which kills apache should the load reach a certain level. This is the only way to stop the “dead-lock”.
I think 2GB RAM would solve the problem (not completely though, because of some rippers), but it takes money ;-)
Comment :: May 10, 2012 at 01:47 :: Quote
That’s indicative of a misconfigured Apache. You should lower the MaxClients/MinSpareServers/MaxSpareServers/ServerLimit and KeepAliveTimeout directives. The solution is to not launch more Apache processes than you have memory for; not rely on disk swap to let you get away with it until there’s some load.
Spiders/rippers shouldn’t be pushing anything out of an opcode cache. I doubt there’s 64MB in PHP code between the blog and the forum, so it should all be cached and your hit rate will be 100% (after the first few minutes).
http://iron.awio.com/apc.php
Comment :: May 10, 2012 at 03:20 :: Quote
“Dan ” wrote:
It is already 40.
KeepAliveTimeout 3
MinSpareServers 1
MaxSpareServers 3
“Dan ” wrote:
It also caches generated code for posts, etc (user cache). So, it’s bigger than 64MB.
BTW, I need to upgrade APC. It’s a bit out of date.
Actually, since I moved to Linode, I can’t complain. There were outages because of overload, but they are quite rare (in comparison to Arvixe).
1 day, 6 hours and 12 minutes
File Cache Information
Cached Files 429 ( 51.4 MBytes)
Hits 10943154
Misses 1013789
Request Rate (hits, misses) 109.92 cache requests/second
Hit Rate 100.60 cache requests/second
Miss Rate 9.32 cache requests/second
Insert Rate 8.08 cache requests/second
Cache full count 3329
User Cache Information
Cached Variables 4125 ( 8.1 MBytes)
Hits 1002465
Misses 680851
Request Rate (hits, misses) 15.47 cache requests/second
Hit Rate 9.22 cache requests/second
Miss Rate 6.26 cache requests/second
Insert Rate 12.43 cache requests/second
Cache full count 1166
Comment :: May 10, 2012 at 03:36 :: Quote