Interested in joining the dynamic and young startup environment of a highly
ambitious Silicon Valley - Istanbul based startup? Keen on learning new stuff,
evolving continuously, experimenting with the latest social technologies? Join
us... GROU.PS' Istanbul office is seeking its exceptional teammates in the
following fields:
- Finance: understand balance sheets? able to execute our well planned
financial projections? willing to communicate with and manage our US and
Turkish accountants? experienced in accounting for more than 2 years? have
background in finance from one of Turkey's top universities? Join us. This is
the place for you to blossom. You may be our future CFO.
- Marketing: understand Google AdSense, AdWords? familiar with landing pages,
A/B tests? keen on experimenting with affiliate marketing plans, coupons, price
discrimination strategies? interested in sales? love social media? GROU.PS is
all about people's passions, affiliations and interests. Join us to execute and
contribute to our well defined growth plans. (http://grou.ps/groudotps/wiki/item/head-of-international-content)
- Customer Support: build a team of interns to help us assist our group founders with live chat and rich documentation. Fluency in English, understanding of social media and good communication/leadership skills are a must. (http://grou.ps/groudotps/wiki/item/director-of-customer-support)
Please send your resumes to hr@groups-inc.com
We are looking for a top level product manager to design and deliver a new
(non-Yonja) major Internet product in the Turkish market. This project (still
in stealth mode) is one of our top priorities after new round of financing from
Tiger Global. Please share this opening broadly in your network. Your help is
much appreciated.
Key attributes we are looking for
- Product marketing / management experience
- Internet / social media domain knowledge or experience
- Excellent communications skills; ability to message and position a product effectively
Key attributes we are looking for
- Product marketing / management experience
- Internet / social media domain knowledge or experience
- Excellent communications skills; ability to message and position a product effectively
- ideally 4-5 years solid experience post university
- these are as important as above: strong team player, self-motivated,
strong work ethic!
- basic conversational English along with fluency in Turkish
The position is based out of our Istanbul office and will be a key leadership role in the company. Competitive salary and company benefits.
To apply for this job: http://www.yonjamedia.com/job.jsp?id=52001
The position is based out of our Istanbul office and will be a key leadership role in the company. Competitive salary and company benefits.
To apply for this job: http://www.yonjamedia.com/job.jsp?id=52001
17 Subat 2011 ulke tarihine gecmesi gereken bir gun, ama gecmeyecek.
Hicbirinizin ruhu bile duymayacak... Bugun Turkiye cok onemli bir degerini
yitirdi, ama malesef kimsenin haberi dahi olmayacak.
Osman Tuncel, Galatasaray Lisesi 133 devresinden cok sevdigim bir
kardesim, arkadasimdi. Malesef onu bugun kaybettik. Su yukardaki paragraf bu
kaybin onemini tanimlamaya yetersiz kalir. Kendisi devrenin en caliskan, en
azimli insaniydi: mukemmel bir matematik-fen yeteneginin yanisira tarihe olan
meraki, caliskanligiyla, daha o yasta (lise caglarimizdan bahsediyorum)
universitelerde tez diye gecebilecek koca bir Kibris arastirmasi kitabini
kaleme almis, etrafinda ben de dahil onlarca kisiyi bu calisma icin organize
etmisti. Cok iyi bir arkadas, dogal bir lider, yardimsever bir insan ve buyuk
bir yetenekti. Vefatinin ardindan devre arkadaslarimin email listesi / Facebook
uzerindeki haykirislari da hep ayni sekilde oldu; Osman seni cenazende degil,
10 sene sonra Turkiye'nin basbakani olarak gormeyi bekliyorduk... Rauf
Denktas'dan, Erdogan Tezic'e kadar 50'li yaslarinda Turkiye'nin bircok onemli
ismi, 18'ine girmemis Osman'imizin bu calismalarini hayranlikla yakindan takip
ediyordu.
Osman akilli bir milliyetciydi; onla olan Kibris sohbetlerimizi
unutamam... Turkiye'ye hizmet etmek istiyordu, bunun icin gerekli donanima
sahipti, bunun icin Turkiye'nin en iyi kaynaklari - lise ve universite olarak -
ona bahsedilmisti. O'nu Kongo'da buyukelcilikteki gorevini surdururken
kaybettik. Hayatin karanliklari onu orda yuttu.
2001 Turkiye'nin en agir ekonomik krizinde mezun olmus, universiteye
baslamis, kelimenin tam anlamiyla buyuk bir enkazi devralmis bir jenerasyonuz.
Ne dogru durust bir egitimi herkes alabildi, ne de universiteden ciktiktan
sonra yeterli istihdam olanaklariyla karsilastik, ne de ise baslanilan yerlerde
cok olumlu bir kultur vardi. Bu sorunlarin bazilari sadece Turkiye'ye ozel de
degil.
Bu dunyanin Osman'a ve Osman gibi "talent" lara ihtiyaci var. Ama malesef
talent burada underrated. Guclu olmali, umidimizi kaybetmemeli ve isigi biz
tutup kendi uzerimize cekmeliyiz. Bence yapacagiz. Hakkini helal et Osman;
basimiz sagolsun.
PS: Email listemize dusen bir video... Osman'i bize hatirlatacak:
After Ning's decision to discontinue their free networks, nowadays I've seen a
lot of blog posts and tweets from community owners, questioning which free
platform to go with. Many of them are comparing solutions such as GROU.PS, Grouply, Spruz, SocialGo, Webs and BuddyPress.
As the founder and CEO of GROU.PS and an avid community manager myself
(http://www.eylem20.org http://www.barcampistanbul.com http://www.socialbeans.org and
of course; http://grou.ps/groudotps)
I'll do something pretty unusual and join this conversation directly.
Here I'll share with you what I think of and know about each of those
do-it-yourself social networking players that are being widely mentioned.
Disclosure: You may think I am biased and you'd probably be right... But my
only purpose here is to be sincere about the market conditions, give you some
insight aside from technical speculation which you can probably find anywhere,
and let you know where my company stands exactly. Yet I still recommend that
you to look at more independent sources as well.
BuddyPress
BuddyPress is a fairly new project but it was created by the same good folks who created the most popular blogging platform, Wordpress. It is a software solution, so unlike hosted/cloud solutions, it requires you to have your own servers and run it from there, continually update and be careful about the conflicts that might occur with your patches and plugins. Which means it's for advanced users and it costs you money. If you are strictly looking for a hosted solution though, I'd recommend that you go with them. The other players are Elgg (which has also recently launched its own hosted solution) and Pligg (which I heard just a few days ago). I know some of the folks behind BuddyPress (at Automattic); they are very nice people and they have funding that makes the software a safe bet for the long-term viability of your community. They are much more experienced in online publishing than any of us, too, but the software is not as mature as that of more established players like Elgg and Pligg. At GROU.PS, we are strong believers in the cloud (i.e. hosted solutions). Running stuff on your own servers costs not only money but time and effort. There's a longer learning curve as well. I'm an open-source addict, and I love how customizable BuddyPress and similar players can be because of their open-source nature. Actually, that's why GROU.PS started its own open-source initiative two years ago. But at the end of the day, we evaluated our resources and decided to stick with one thing—but do it so perfectly that we would deliver the best of both worlds. At GROU.PS, we have the vision to create an eco-system similar to -- or even broader than -- SalesForce's AppForce platform. Our Developer and Template Design platforms are early examples of that. Expect to see tighter integration and more opportunities to make money from your knowledge on the GROU.PS platform soon. Such an extensible cloud-based solution not only is much more practical but would solve many of the incompatibility and crash hassles that occur in the open-source world.
Spruz
[update: Spruz has brought new limits to its service, only 100 members allowed in the free package]
Spruz has done some pretty impressive stuff. It managed to sneak into the Ning network with a toolbar add-on called skysa, and then created its own alternative platform, which today is called Spruz (and looks like a quick and dirty hack), using the knowledge and expertise it gained from skysa to steal networks from Ning. This strategy gave it a huge boost over others in the importing process. Independent stats confirm that, too: Alexa shows stunning growth, and Compete and Quantcast indicate that GROU.PS and Spruz are the fastest-growing networks since Ning’s big announcement. This growth of Spruz is fascinating, considering that GROU.PS had already been Ning’s biggest competitor for a long while; showing the same growth as us as a new player is indeed a good job. Spruz remains the fourth player, though, right behind SocialGo.
GROU.PS
According to all independent sources GROU.PS is today the largest competitor of Ning. And again, early results show that GROU.PS and Spruz have become the biggest winners of Ning's decision to kick off free networks (even though GROU.PS has no leverage such as skysa) But I think there's much to see in this area; so everything can change. What I can observe people create their networks to try the platforms but many haven't come to a final conclusion yet. And I think that makes sense. Platform choice is a very important decision and needs to be made carefully.
GROU.PS is a content agnostic platform that aims to monetize this space by an asymmetric business model. That means, unlike Ning, we won't sell software, that's not our plan, we will instead charge brands and advertisers. Asymmetric model works pretty well in Google, Facebook and Twitter. On the other hand, services like eBay, Amazon, Salesforce use a symmetric model (a la Ning) where they charge directly those who use their service.
Final Notes
Like this article?
Please retweet, digg, delicious, Facebook and/or share this with your friends on Ning.
We care about our communities. After observing a lot of education communities moving to our platform; we've decided to create 2 new templates for the education community. Here are the screenshots:
BuddyPress
BuddyPress is a fairly new project but it was created by the same good folks who created the most popular blogging platform, Wordpress. It is a software solution, so unlike hosted/cloud solutions, it requires you to have your own servers and run it from there, continually update and be careful about the conflicts that might occur with your patches and plugins. Which means it's for advanced users and it costs you money. If you are strictly looking for a hosted solution though, I'd recommend that you go with them. The other players are Elgg (which has also recently launched its own hosted solution) and Pligg (which I heard just a few days ago). I know some of the folks behind BuddyPress (at Automattic); they are very nice people and they have funding that makes the software a safe bet for the long-term viability of your community. They are much more experienced in online publishing than any of us, too, but the software is not as mature as that of more established players like Elgg and Pligg. At GROU.PS, we are strong believers in the cloud (i.e. hosted solutions). Running stuff on your own servers costs not only money but time and effort. There's a longer learning curve as well. I'm an open-source addict, and I love how customizable BuddyPress and similar players can be because of their open-source nature. Actually, that's why GROU.PS started its own open-source initiative two years ago. But at the end of the day, we evaluated our resources and decided to stick with one thing—but do it so perfectly that we would deliver the best of both worlds. At GROU.PS, we have the vision to create an eco-system similar to -- or even broader than -- SalesForce's AppForce platform. Our Developer and Template Design platforms are early examples of that. Expect to see tighter integration and more opportunities to make money from your knowledge on the GROU.PS platform soon. Such an extensible cloud-based solution not only is much more practical but would solve many of the incompatibility and crash hassles that occur in the open-source world.
Spruz
[update: Spruz has brought new limits to its service, only 100 members allowed in the free package]
Spruz has done some pretty impressive stuff. It managed to sneak into the Ning network with a toolbar add-on called skysa, and then created its own alternative platform, which today is called Spruz (and looks like a quick and dirty hack), using the knowledge and expertise it gained from skysa to steal networks from Ning. This strategy gave it a huge boost over others in the importing process. Independent stats confirm that, too: Alexa shows stunning growth, and Compete and Quantcast indicate that GROU.PS and Spruz are the fastest-growing networks since Ning’s big announcement. This growth of Spruz is fascinating, considering that GROU.PS had already been Ning’s biggest competitor for a long while; showing the same growth as us as a new player is indeed a good job. Spruz remains the fourth player, though, right behind SocialGo.
I really dislike a few things about Spruz, though, and I should state them here:
1) The look and feel are pretty poor, although the templates are fairly customizable. I know that style is a subjective matter, but try it for yourself. Look and feel are critical to creating a sense of belonging within your online community. They are what makes your members feel at home. Otherwise, why not use Facebook Groups to start your own social network?
2) Spruz is new but is already starting to have
technical issues. Scaling to thousands of social networks is not an easy job.
We remember that from our early days in 2008. Scaling requires a lot of
experience and investment.
3) A personal concern of mine is that some of its templates are exact copies
from our library. Granted, some of those templates are open source, and
imitation is OK to some extent—but copying our default template is not the
friendliest thing you can do.
4) One thing that I recommend all community managers look at is the financial
stability of the platform they choose. Hosting millions of networks is an
expensive job. Spruz is secretive about its funding and corporate status.
Choosing a platform is like choosing a business partner: you need some
transparency.
5) In a blog post, Spruz declared that it would cut costs by removing unused communities and by not serving videos. Not serving videos or files and saving on bandwidth is understandable; bandwidth is expensive. But losing that functionality is serious. And removing unused communities is kind of a silly argument because that does not really cut many costs.
Apart from these points, I find Spruz' community
management very successful. It seems very responsive to people who want to
build their communities there.
[Spruz' response can be found in
comments]
SocialGo
I first heard about SocialGo, I guess, in mid 2008. Their platform is noticeably different for those who are familiar Ning. You may consider this originality and hence a plus.They are surely spending the most money in between these companies for advertising. I even see their ads on GROU.PS. They invest a lot of money in Google Ads. They should be financially more secure compared to some of the other hosted DIY social network companies that I mention here; their investors include the founders of Eidos, a big video game success.
SocialGo
I first heard about SocialGo, I guess, in mid 2008. Their platform is noticeably different for those who are familiar Ning. You may consider this originality and hence a plus.They are surely spending the most money in between these companies for advertising. I even see their ads on GROU.PS. They invest a lot of money in Google Ads. They should be financially more secure compared to some of the other hosted DIY social network companies that I mention here; their investors include the founders of Eidos, a big video game success.
But the current results show that there's not much traction on SocialGo, people
tend to use other platforms instead. A personal note, it never appealed to my
tastes either. And perhaps the fact that almost every single thing in their
admin panel is a premium feature must be annoying for the group founders.
Webs
Webs is a very old and pretty stable web site creation service. I know one of their founders and investors, they are pretty committed to what they do and they are right now one of the best web site creation services out there. However, DIY social networking is not their core; it's an after-thought. Their DIY social networking platform is built on top of their existing web site creation platform as a patch - and that shows in the product itself with very weak social graph features, no real benefits of friending/following etc. That's why DIY social networking sounds like a me-too thing that they've done to justify their valuation in the heydays of the social networking.
Webs
Webs is a very old and pretty stable web site creation service. I know one of their founders and investors, they are pretty committed to what they do and they are right now one of the best web site creation services out there. However, DIY social networking is not their core; it's an after-thought. Their DIY social networking platform is built on top of their existing web site creation platform as a patch - and that shows in the product itself with very weak social graph features, no real benefits of friending/following etc. That's why DIY social networking sounds like a me-too thing that they've done to justify their valuation in the heydays of the social networking.
All the tweets and blog entries I come across show weak signs of interest into
Webs. I think that's a result of the weak social graph features I've just
mentioned. People, when they sign up, realize they're not on an online
community but rather a web site creation service.
Grouply
Grouply has a special place for me. 3 years ago, when Grouply was about to launch, GROU.PS was already up and running, I was introduced to them by my advisor who told me that we might be doing something similar - probably concluding from our names. Anyways, I emailed back and forth with their founders, I always enjoy meeting the managers/founders of the competitive companies, compare notes and see what we can do together to create a healthy competition environment. At that time, GROU.PS was totally a 1 man show with a single advisor, I had no funding, I was just a 23 years old guy who left his homeland with the dream of creating the world's best online community in Silicon Valley - where the software dreams come true.
Grouply
Grouply has a special place for me. 3 years ago, when Grouply was about to launch, GROU.PS was already up and running, I was introduced to them by my advisor who told me that we might be doing something similar - probably concluding from our names. Anyways, I emailed back and forth with their founders, I always enjoy meeting the managers/founders of the competitive companies, compare notes and see what we can do together to create a healthy competition environment. At that time, GROU.PS was totally a 1 man show with a single advisor, I had no funding, I was just a 23 years old guy who left his homeland with the dream of creating the world's best online community in Silicon Valley - where the software dreams come true.
However probably this extremely idealistic character of mine made Grouply
founders consider me as a potential hiring candidate rather than a competitive
service provider - which I wasn't OK with of course. Months later they've
launched a service which was pretty different than DIY social networking, it
was an aggregator for Yahoo Group and Google Groups. But after receiving a huge
number of spam complaints, it didn't take too long for them to change their
product. Just last year, they came up with a DIY social networking offering,
which is of course, in terms of product, years behind what we offer; as a
result of starting being pretty late to the game.
I can see that Ning's decision pushed some community managers to try Grouply as
well, but the traction doesn't seem large yet.
GROU.PS
According to all independent sources GROU.PS is today the largest competitor of Ning. And again, early results show that GROU.PS and Spruz have become the biggest winners of Ning's decision to kick off free networks (even though GROU.PS has no leverage such as skysa) But I think there's much to see in this area; so everything can change. What I can observe people create their networks to try the platforms but many haven't come to a final conclusion yet. And I think that makes sense. Platform choice is a very important decision and needs to be made carefully.
GROU.PS is a content agnostic platform that aims to monetize this space by an asymmetric business model. That means, unlike Ning, we won't sell software, that's not our plan, we will instead charge brands and advertisers. Asymmetric model works pretty well in Google, Facebook and Twitter. On the other hand, services like eBay, Amazon, Salesforce use a symmetric model (a la Ning) where they charge directly those who use their service.
We also believe in revenue share. We know you, as an organic content creator,
usually are not motivated by money to create that content and get people
around. But you'd love to see the benefits of your efforts and/or create a
budget for your community. GROU.PS aims to do that (and already does it to a
some extent) and cut a commission to pay the bills.
This kind of a business model requires us to be very careful about our costs
and there comes our competitive advantages. GROU.PS' development team is pretty
large yet it is based in Turkey. That allows us to create a great product at a
much lower cost than our competitors. As for bandwidth costs (GROU.PS provides
enormous amounts of bandwidth and storage space for its users) we have a
stealth mode product that we will announce soon - and you'll see how we're able
to keep the things going by some truly innovative methods. I think you should
definitely keep those financial arguments in mind while choosing your platform
of choice.
People usually criticize GROU.PS for not being stable and I understand that.
Indeed we did have periods of time when the service was on and off and our
early adopters showed great patience - which we really appreciate and will
never forget (even in our pricing) But we gained so much experience at that
time that I am proud to say, today we are operationally very very stable. As
for the product stability, we're getting better at it too. While having a team
based in Turkey is a competitive advantage in terms of cost management, it was
also a drawback in terms of software engineering experience. Nobody can beat
Silicon Valley's experience in software engineering. But we have an office in
the heart of Silicon Valley too, and that helped our small team quickly gain
the same experience as their peers in Silicon Valley and today we have a very
talented/gifted team and software engineering methods that allow us to deliver
you a very stable product. However, we still keep the beta tag in our logo,
because our product is not at the place where it needs to be yet. Our product
vision continues to drive us further and further, stay with us and you'll see
some very cool stuff coming up. Of course, some rare instances of instabilities
are possible due to continuous release and other agile methods that
we follow - but we'll keep that very very low.
I think another thing that I need to discuss is funding. GROU.PS has done a lot
with very little funding. But, although I'm not able to get into details yet,
you should be confident that we're about to announce something soon that will
give us a long runway to accomplish the big vision that I have very vaguely
outlined here.
All in all GROU.PS is like Japanese cars. these are our baby steps, you see a
great product, great innovation but you hear about some instabilities; but
that's about to change; GROU.PS does not just keep innovating but is also
creating a truly amazing product with superior stability than anybody else. We
are committed to keep our business model asymmetric (which means free for our
group founders) and innovate in terms of business as well.
Final Notes
Well I know that this post has been in favour of my company, but that's because
I truly believe in everything I say here. I invite other CEOs/product managers
to post their own thoughts in comments or as new blog posts and I promise to
publish links to their posts from here - for the sake of objectivity and giving
a very well deserved say to them too
I anticipate that Ning's decision to kick off their free networks will result
in many international networks to leave. That will make Ning more US focused.
But I believe, people, in time, will choose the best product in the market - no
matter what, no matter where they are. Gina Bianchini, ex CEO of Ning (and a
terrific person whom I had a chance to meet), once said in an interview that
she believes Ning is gonna remain number 1 in this space because the network
effects will speak and keep the status-quo - her examples were Facebook and
Twitter's points of displacement. With all due respect, I don't agree with
that. First, DIY social networks consist of networks that are isolated from
each other, network effects take place within our micro social networks not
through the entire network itself. Second, even though the network effects were
to take place, don't forget that Friendster and MySpace were number one before
Facebook took over the throne.
Like this article?
Please retweet, digg, delicious, Facebook and/or share this with your friends on Ning.
BONUS MATERIAL:
Feature by Feature Comparison Sheet
GROU.PS Growth Figures
We care about our communities. After observing a lot of education communities moving to our platform; we've decided to create 2 new templates for the education community. Here are the screenshots:
Google has unveiled its upcoming browser based operating system, Chrome OS. If
you haven't seen it yet, here's a video: (via Webrazzi)
It is pretty similar to what I described 3 years ago on ReadWriteWeb:
GoogleOS: What to Expect. In this article, I
was proposing 3 scenarios, but my pick was the ByzantineOS one, which pretty
much resembles to what we see today.
I'm eager to download and give it a try. Did anyone see it live, is
dual boot hassle-free? I don't think I can give up Windows right away - which
also makes me think Google may either offer a Windows in the cloud service (or
let Microsoft do it via Azure) or acquire virtualization technology
inside Chrome OS.
Google has a great marketing campaign that is now on the front page of Google
Docs; it's called Google Docs, a Love Letter and here is their video:
My natural first reaction was to share it with my friends via Twitter and
Facebook, it's so funny. And this is how you capture market share from the
incumbent Microsoft, do something that the new entrants, youth, will love and
spread virally.
And that's not all for me. I have also taken time to pause the video and
write down what Michael was about to send Jessica before his friends came to
rescue. Here is his amazing art work, totally made my day:
Dear Jessica,
To quote our old friend William Shakespeare:
"See, how she leans her cheek upon her hand! O that I were a glove upon that hand, that I might touch that cheek!"
From the first day I saw you in French 101, I only wanted to be that metaphorical glove. Though my first instinct was to tell you "je te adore," I now long to say, "Je veux etre avec toi!!!" I know we haven't talk much thse past three years, but I know you are
*~*~* The One *~*~*
Based on this conclusion, I am making a proposition to which I am hoping you will consent. In order to satrt our courtship, it is first necessary to meet in person and complete the procedure of a proper date. This could include any of the following proposed activities (though not necessarily in the current order). I have highlighted my preferred choice:
DATE ACTIVITIESPlease let me know in a prompt fashion whether you will partake in this date with me. An idea of your schedule in the next week would be helpful in planning this event. Please try to avoid Wednesday and Friday nights, as that is when my dorm has game nights, although you are free to stop by to join in the festivities if you would like. Sorry, that came out wrong--actually, I would love it if you came to those too, that is, if you want. Would you mind responding by tomorrow?
- A film at the local theater
- A beverage of some sort
- A brief stroll through downtown
- A visit to the mini gold establishment down the street
- Playing any of an assortment of MMORPGs
- A tour of the Computer History Museum
Dreaming of you,
Michael
Turkiye'deki internet yasaklari sizin de mi caninizi sikmakta.
Siz de bunu bir ozgurlukler meselesi olarak goruyorsaniz ve benim gibi bu
olanlara sessiz kalamiyorsaniz buradan buyrun: www.eylem20.org
Biz sosyal medyanin gucunu kullanarak bir taban hareketiyle sesimizi
tavana duyurabilecegimizi, Turkiye'de bazi seyleri olumlu yonde
degistirebilecegimize inandik.
Cok gec degil, birkac gun once Google'a verilen hakli sayilabilecek vergi
cezasi dahi, internet yasaklari konusunda yaptigimiz hatalar yuzunden, onumuze
getirildi ve konu siyasi bir hal aldi; uluslararasi arenada elimizi
gucsuzlestirecek bir diger konu oluverdi.
Siz de bu tarz banner'lar web sitenizde kullanarak (http://www.eylem20.org/36853) eylem'e
yardimci olabilirsiniz. Ayrica uye olmaniz, arkadaslarinizi cagirmaniz
(Facebook Connect ile), tweet etmeniz de gayet kafi olacaktir. Hemen su alttaki
butona klikleyerek baslayabilirsiniz.
Things to bing (or google if you want to):
nginx, libevent, memcache, danga interactive, gearman, mogilefs, nosql, cassandra, hive, hadoop, scribe, thrift, maatkit (for mysql), mmm (for mysql)
MySQL useful settings:
query_cache_size = 0 # on master
query_cache_type = 0 # on master
thread_concurrency = 8 # total cores
max_connections = 750 # shouldn’t exceed that
innodb_buffer_pool_size = 10G # a little less than the total amount
And our typical sysctl additions; as I've promised - the configuration that lets us serve 1PB per month to 3 million unique visitors:
net.ipv4.tcp_syncookies
= 1
net.ipv4.tcp_synack_retries
= 2
##
Emre
edited
#
http://www.oracle-base.com/articles/11g/OracleDB11gR1InstallationOnFedora8.php
kernel.shmall
=
2097152
kernel.shmmax
=
2147483648
kernel.shmmni
=
4096
#
semaphores: semmsl,
semmns,
semopm,
semmni
kernel.sem =
250 32000 100 128
net.ipv4.ip_local_port_range
= 1024 65000
net.core.rmem_default=4194304
#net.core.rmem_max=4194304
net.core.wmem_default=262144
#net.core.wmem_max=262144
fs.file-max=5049800
vm.swappiness=10
##
Emre
edited
# from
http://forums.softlayer.com/showthread.php?t=3252
net.ipv4.tcp_rmem = 4096
87380 8388608
net.ipv4.tcp_wmem = 4096
87380 8388608
net.core.rmem_max
=
8388608
net.core.wmem_max
=
8388608
net.core.netdev_max_backlog
=
5000
net.ipv4.tcp_window_scaling
= 1
net.ipv4.ip_nonlocal_bind=1
#
http://rackerhacker.com/2007/08/24/apache-no-space-left-on-device-couldnt-create-accept-lock/
kernel.msgmni
=
1024
kernel.sem =
250 256000 32 1024
net.ipv4.ip_conntrack_max
= 524288
net.ipv4.netfilter.ip_conntrack_max
= 524288
