Superjedi

Has anyone been running whmcs and percona cluster on arm64 servers instead of the usual x64 architecture? How is the reliability in general? Is it solid?

I found the problem: the stripe module with missing passwords...

I just updated to 7.8.2 and if I try to check out after ordering the product, there is the error message that I cant get rid of. Any idea which it comes from? Many thanks.

he doesn't even reply anymore. That is a bad sign that he is stuck.

Will there be an official support for connecting to MySQL 8 without running a script inside MySQL 8 to go back to the old login method? Whmcs and WordPress have the same login incompatibility with MySQL 8. https://dev.mysql.com/doc/refman/8.0/en/caching-sha2-pluggable-authentication.html https://github.com/docker-library/wordpress/issues/313

Correct. The certificate is valid and it is provided by cloudlfare. In order to get back into the admin, I have to go into MySQL and change whmcs system URL back to http://example.com. That is if someone is kind enough to shed some light where whmcs is determining if my site has SSL protection ? There has to be a logic check.

I am running whmcs with php/apache/mysql in docker. I also use cloudflare for DDOS protection and letsencrypt from caddy. Even though the site has been in https with ssl from cloudflare, the admin warning says my site isnt ssl protected. The main issue is after I change the WHMCS system url from http://example.com to https://example.com, the site gets blocked out. I have checked every post related to this issue e.g. using curl -I https://example and the output is 200 so I cant see why the admin ssl says I am not using ssl. these dont help: https://help.whmcs.com/m/troubleshooting/l/802231-troubleshooting-a-redirect-loop https://help.whmcs.com/m/64969/l/680976-updating-the-system-url https://whmcs.community/topic/268037-system-health-status-website-ssl-warning/ I have a feeling that if I dont get this fixed, I will encounter other ssl related issues later on e.g. other payment plugins refuse to work because whmcs thinks I am not using ssl even though the site is definitely ssl protected. What is the whmcs system logic in determining my site isnt ssl? I have checked using curl -I https://example.com and the result is 200 so warning logic isnt based on it. Thanks in advance for any insight? Vince

In order for the webhook of whmcs to disconnect or connect a VPN node based on customers payment and usage , the download traffic is monitored. Tunnelbear tracks the download traffic in real time so they can disconnect a customer easily if a customer has exceeded his allowance . How does it affect the performance of whmcs database instance if it has to track all the VPN traffic in real time when there are 1000 nodes streaming and serving 10000 customers ? Can I assume the whmcs database will aggregate all the usage data from different VPN nodes for each customer ? For example, I have to set up and erase each VPN node frequently so any data on each VPN node will be erased and if whmcs database doesn't record and save the traffic log in real time and at the same time I delete some VPN nodes when the whmcs database is spinned down for maintenance, those traffic log will be gone. Another scenario is that if I have to spin down my whmcs main database instance for maintenance, all the vpn will still be operational but the real time download traffic wouldnt be processed by the whmcs database. Will the whmcs database sync up to those download traffic log of each VPN node when it is spinned back up again ? Or those traffic database will be gone ? These are quite important issues because it affects how we do maintenance and the integrity of the traffic log wouldnt be ensured depending how the database tracks the traffic..

Ibvpn also allows users to manually select the server to connect in addition to assigning automatically the quickest server for connection. For most commercial VPN providers, they won't let user select each node manually because if a node is too full, it may degrade the speed of all the current users in that node further. If a bus is already too , u don't want more passengers barging in. E.g. tunnelbear only assigns a node based on geolocation. I think they just use a simple highest speed criterion for the allocation for each region. Their allocation and connection is instantaneous unlike ibvpn's waiting for least 30second. Since I am not ibvpn, I wouldn't know why it takes so long to assign the quickest server for a connection. My best guess is that they just query the entire list of server for the highest connection in real time and when they only have 20 server on the list, it is quick. But when their list grows to 175, the query takes 30s to run. Obviously they couldn't find a way around this slow connection problem. In my opinion, that is an unacceptable waiting time because if they grow to 1000 server, the auto connection time will be 2minute. So the big question is whether u should list every VPN node config for users to choose manually. This is normal for only small vpn business because each region wouldn't have more than 20 servers and so users can eyeball the fastest one to join. Ideally and if it is possible to query dynamically and in real time without delay , user should only be shown the quickest vpn node for each country. The best scenario, a VPN operator should be given warning if all the servers become slow and it is time to spin up new VPN nodes to cope with more traffic. I personally use terraform and docker to spin up each node on the cloud and the bottleneck of setting up the node is the manual product input using whmcs GUI. If you want to discuss more in details, feel free to contact me in telegram @Superjedi.

I have found one vpn company which has done a good job in using whmcs well. Check out ibvpn.com. however their client logon mechanism is very slow because they have to query the entire list of vpn nodes around 170 of them before they allocate the best one based on least bandwidth% or user number. It doesn't look like a very scalable design.

Does this addon keep track of each VPN node bandwidth usage % and the number of users per node in real time ? It is important to have the logic to assign the 'least busy' node automatically to a customer for connection.

Most vpn providers provide vpn based on geographical locations e.g. UK, USA , Japan , France etc so users don't select each node within each country themselves e.g. there may be 20 nodes in USA , 5 nodes in Japan. Users only select the location and they don't actually see all 25 nodes. How will whmcs addon allocate the users among the nodes in each country ? I ask this question because my current whmcs add-on gives a list of nodes to be chosen by the user e.g. he has to select node 10 in USA or node 2 in Japan and if node 10 in USA is overloaded and slow , he will have to try node 7 in USA or node 11 in USA manually until he finds one which is less congested. So if there are 100 nodes in USA and 99 of them are very congested , the user might have to try 99 nodes before he can find one that is not congested. How will this situation be handled with wireguard ? Is there any automatic node allocation or it will be a manual selection process ? For example if there are 100 nodes in USA , 50 in UK , and 200 in germany, user has to go through a list of 350 nodes manually in order to log into a node?

Is the module ready for beta testing soon ?

I am currently using shadowsocksr and v2ray in my service as there are whmcs plugins for multiuser management similar to your project. U should check out those open sourced plugins as they work quite well. Their design and structure may help u shorten building time.

Great. Would it be ready within 3 months ? Or more like around 1 more year? I would be more than happy to be an early beta tester.

When will the wireguard version be ready for beta production to download and operate ?

Would it be possible to Create a wireguard version ?