I am having a hard time choose (runcloud vs. enhance): Page 3 - Enhance Control Panel

I am having a hard time choose (runcloud vs. enhance)

gmakhs

slimx @Adam Seems that many of us have issue with the auto fix 503 missing, can something be done about it?

sikmo

slimx I asked in support ticket and and as far as I understand, this is expected behaviour. Purely based on website consuming too much resources. And the website should be optimised.

gmakhs

sikmo 503 can simply xaused by a ddos or dos attack, or just normal high traffic its not only optimization related.
Auto fix 503 help you get the headache out of such issues and makes sure the website operates while the customer decides what to do instead fo showing 503 for hours.

not having it is not expected neither good

Zoinkies

90% of our support tickets from our customers are 503 issues.

gmakhs

Zoinkies i hope @Adam really takes it seriously and come with a solution, after all we purchased litespeed to be able to use all its features and enhance advertises that it works with litespeed so it should support all features

slimx

What's annoying is.. naturally like im sure most of you.. I have Openlitespeed on the "lower-tier" servers.
I put LiteSpeed Enterprise on my "good servers". It's annoying that the server with more juice, is offering less...

Especially that annoying "hang" that it does when it hits that 503. It doesn't even take you to the page, it has that 5 minute hang time lol.

Adam

503 errors are caused by the website hitting its nproc or memory setting, either because it's very busy and requires more PHP workers than the assigned resources permit, or PHP workers are building up due to I/O or CPU limitations or long running database queries,

As far as I'm aware, the "auto fix 503" feature in Litespeed restarts PHP to kill any running processes - this would serve a 502/503 to existing visitors but in theory allow new requests to succeed. The reason this doesn't work in Enhance is because PHP is started within the website container and isn't managed by Litespeed. It's not a bug - built in containerisation is one of the core features of Enhance.

We could re-implement the "auto fix 503" feature in Enhance but feedback from our mass hosting partners is that this behaviour isn't desirable - on shared hosting the customer should upgrade their plan or make their site more efficient and on dedicated hosting the limits should be set sufficiently high in the first place.

I'm not aware of any difference in PHP behaviour between Openlitespeed and Litespeed.

In the case of one poster in this thread, based on previous support tickets, they have no memory/nproc limits whatsoever and their server is going out of memory. This isn't something Enhance can mitigate.

sikmo

Adam I will have to disagree on this, but as I mentioned, we are still in testing and will bring forward the results, whatever they are.

One of such cases - updating plugins in middle of night without any visitors on the website. I hit the NPROC and RAM extremely fast, which is not something I can't replicate anywhere else - and I am comparing behaviour of OLS/LSE with Enhance to OLS with RunCloud (so anyone can test it for themselves).

Note: NPROC limit is set to 24. with 3 GB RAM. Since each process can have ~ 200 MB size, you can see how fast you can hit the limit. And that would not be the issue, if processes were swiftly removed - but they are not. Which is currently our main investigation point.

Note 2: If NPROC is set to 24 with 6 GB RAM, the 503 are not as often. As the max theoretical sum of 250 MB per process multipled by 24 is 6 GB. But we've had issues with websites hitting 2 GB RAM with 24 NPROC in middle of night with almost zero visitors (at that moment I believe only CRON was running and doing nothing).

We've tested lowering numbers such RetryTimout and others with clear results mitigating the 503 errors with success. Processing hanging there for long periods of time instead of ending themselves is seem to be the root cause for most of our problems.

davemcl

I use Plesk, RC & Enhance - currently leaning on Enhance to replace Plesk. RC I primarily use for managed Wordpress and Laravel sites.

Adam

sikmo

lsphp shuts down idle workers after 300 seconds (5 minutes). If you're running your cron jobs with PHP CLI then this will exit immediately and leave no worker behind.

To see which limit you're hitting you can look at /sys/fs/cgroup/websites/{websiteId}/pids.events and /sys/fs/cgroup/websites/{websiteId}/memory.events to see which limit you're hitting.

One thing to bear in mind, the nproc limit in v2 cgroups counts o/s threads. Therefore if your PHP application spawns threads or forks processes then this might be the cause of the nproc limit being reached.

gmakhs

Adam A small DoS attack can make any website hit 503 , or website hitting google discovery , even small mistakes during dev or updates can, the customer ends with 503 errors until restart container instead of the website showing temporary 503 errors, if your big hosting client don't like it its ok, and i respect that this is why it needs to be OPTIONAL the same way is on litespeed, same goes for the file manager.

P.S also something indeed is wrong with the way PHP works on litespeed , there are issues with hanging that cant be replicated on other panels

Zoinkies

Yea there is 100% something wrong with Litespeed on Enhance. We’ve had to move customers we actually care about to less powerful VPS but using cPanel, instead of enhance. No 503 issues post move, enhance however has 503 issues I can’t narrow down. I bought the unlimited worker litespeed enterprise license, 503, bought a 16c core 6Ghz server with 256GB RAM, 503, set unlimited package restrictions 503, nothing I can do fixed it. Doesn’t happen at all with enhance?

gmakhs

Zoinkies ITs the main reason i haven't manage to migrate not even a single client from our DA servers and we use enhance only for inhouse projects and some chosen clients only ,we get no issue on DA - CL and the same website even on a more powerful server and with more resources throws 503 on enhance

Thomas_W

gmakhs Enhance file manager is bad specially if server have over 100 ms difference between them (control and slave server ) same goes for phpmyadmin which is nightmare to work with

I have a server with 180-190ms latency. The file manager makes for me no issue at all. But phpmyadmin can not be used. I setup there a phpmyadmin at a domain for customers with some security. Of course it's not a single login, but the customers are fine with this solution.

For the file manager is maybe also the disks and server speed a critical point. I have fast servers.

gmakhs

Thomas_W file handling is ok , but it times out , which results to uploads failing or stuck , sometimes it disconnected you if you stay long term while updating files and testing , it's not hardware issue , server are fast and empty

Thomas_W

gmakhs I work there at one project also sometimes also a longer time with the file manager. But there is also a customer request for uncommon high sessions for PHP. Maybe try to put session.cookie_lifetime and session.gc_maxlifetime up. Or ask Adam, I don't think it's the latency.

« Previous Page