I made some updates to the server's PHP-FPM configuration using the advise of a site to optimize PHP_FPM. So far so well. I have noticed better performance and stability and I am hoping that at this point the Instance is stable enough to handle the mild pressure it receives from its few users.

I am also working through a list of server optimizations for Friendica. Some I already had in play others needed tweaks. Stability is my main priority. I have limited the number of Users to 100. I will increase this if we reach 100 users and the Instance handles the load well.

This is a small Instance intended for family and friends to communicate. This Instance is not for Marketing your services or products. I do not mind if you mention your product or service though you may not make constant mention of it spamming other users.

I do not have the resources for a person to effectively launch a marketing program from my Instance. Not to mention the Instances propagation is not what you would desire for marketing. If you want to market on Friendica I recommend going to a larger Instance. It will not be allowed here.

Please refer to the TOS & Rules regarding this behavior

The Instance went down again. Instead of rushing to get the instance back up I check some logs and it appears that when I increased the values I did not increase the pm.max_spare_servers. I allowed more to be generated but did not allow them to exist 😉. Which of course took php-fpm down.

I increased pm.max_spare_server so hopefully this time I have enough to handle the load. As you can see in the screenshot below as soon as more spare servers were added that brought it above 100, the whole service shutdown bringing the site down.

The setting still may need massaging. Time will show.

I was on the Instance off and on last night at work on breaks. When I got home I noticed the site was down again due to a HTTP 500 Error. I have done some more massaging to the php-fpm config though I will need to get some rack before I can dig into it and find out what optimum setting would be for my server giving its memory.

Sorry for any inconveniences still trying to tune the server. I am not sure at all how long the server has been down.

I went in and reconfigured my php-fpm pm.start_servers values as it appears default pool was too small and would get busy quite quickly. I am hoping this cures the issue. I found no other PHP errors to deal with so I hope this is the culprit. So I am hoping that getting this fine tuned and all will be well on the Instance, or so I hope. 😀

note: The default is really low which is 50. If I can exceed that limit with only 3 active users I am sure it will have to be much higher. I will have to do more research to see what a good value should be.

I woke up today to find my Instance was down. I am not sure why at this point. It is a work day today and I will be busy this weekend so not sure when I will be able to investigate further as to why this is happening. I have my suspensions though there will be no easy fixes for this I am afraid.

note: The VPS was still up and running so it was not the issue. Restarting httpd & MariaDB only brought it from a 500 to a 503. So I will have to rummage through the logs to see what is going on.

note: Some preliminary research shows this is going to be a PHP related issue. Which is good and bad. As it means it is code base, as I have a pristine code base (no alterations) this is an upstream issue. It may be in an Addon? That would be nice disable the Addon problem fixed. Though I am not often that lucky. I will look through the PHP logs and see which process jammed the system 😉.

Sorry for the inconvenience this has caused.

I am not sure how long the Server has been down today. I went to login and got a 503: Service Unavailable. I am looking into what caused the issue.

After discovering that I could not connect to my instance, though the server was still up. I did the most reasonable thing ;). I rebooted it. That seems to have fixed whatever took the service down. I am looking into what caused the issue and how to prevent it in the future.

I have managed to fine tune my instance for now. Though I will have to see how it handles heavier loads when new users signup. For now the logs only show anticipated issues. Such as graphics that are shared that are not to encoded to the standard and thus throw exceptions. I would rather those exceptions be handled though for now it is not a big deal as it does not affect the server performance.

I am still learning the Friendica codebase so it will be sometime before I can work out issues such as these. And there are far more pressing issues.

Okay, Rogue Project Friendica is setup and tuned as well as I can with no user base. At this point I will just wait to see if I get some users and see how the server behaves.

libpng has been throwing some warnings though that is typical with a library that handles so many files created from other resources, that are not 100% compliant to the standard. So far it has not caused an issue. I will be keeping an eye on it. 😁

I had a glitch earlier tonight which caused me to have to drop my database and re-install. Fortunately it was a fresh install and nothing except test data was lost.

Rogue Project: Friendica instance is now, once again in Open Registration.

Keep in mind it is a new instance and there will be some tuning to do.