New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Having the same issues as well.
Currently on LV-02 Reseller and getting downtime almost every week. https://imgur.com/a/YS17R
Litespeed is getting hard reset, not sure what to do on that but wait for a reply.
Litespeed at one point did apply an update for it in 5.1.11:
" Fixed an issue where cPanel killed LiteSpeed when applying configuration changes, such as adding an add-on domain or sub domain."
It's possible its the same bug again.
Francisco
I can respect that and the honesty in your responses. I appreciate you taking the time to address my concerns and suggestions with such a meaningful dialog. From the responses here I can now better understand the direction buy shared is taking and believe as you've said that it'll continue to improve and stabilize. Thanks for your efforts and service. P.S. twice the price with better stability for plans like the 1gb and 5gb would be of interest.
Ya I hear ya. I actually have an established and reliable VPS provider that can offer solid pricing to me. But I'm currently too lazy to setup, lock down, maintain, and backup another VPS. But I'm also bit hard pressed to pay VPS prices for shared hosting. IMO a KVM is always better than shared hosting. Provided your not too lazy or incompetent to set one up.
Maybe that's what it comes down to if you want really reliable shared hosting. Hosted databases (think Amazon RDS) are almost always more expensive than self-hosting since you're paying someone else to handle all the drama and fire drills when something happens. Web servers aren't really that much different.
I have a 1gb buyshared plan at $5/y (with a dedicated ip even!) and at that price I figure I can't expect much. I'm not bothered by occasional blips as long as there is some basic monitoring and not too many extended outages. I've mentioned a few times that I know people who would like a more expensive, high availability tier and I think BuyVM is in a good position to offer that and it would be great if they did.
I have a family member with a small affiliate blog who was itching to pay $20/month for HA wordpress hosting (she had lost significant revenue at other hosts in Black Friday outages, which are the highest-income but most crash-prone time of year). That would be an awesome BuyVM product if you ask me. But HA means real HA, i.e. the person's site is automatically replicated across multiple DC's with an anycast address aiming at zero downtime. I still like to (wishfully?) think that the economics can work out.
My web site is down, I'm losing millions!!!
There's a good chance that the hard restarts we see on litespeed are intentional by cPanel and it's just Litespeed takes a few to process things. If I threw some 3.X GHz cores in there it may very well smooth that right out.
I've been eyeing some E5 2667's as good contenders for the new processors. I got 2660v2's in lv-shared03 and they're good, but a nice high single thread bump would be a good pickup.
Francisco
Cross DC HA is a nightmare with SQL unless I have 2 DC's in the same city and just have a loop between them.
We'll see what block storage does for us. If it performs like I want it to we could very well look at offering some sort of HA options inside Stallion besides just IP floating.
Francisco
Cross-DC block storage?! :O
I thought cross-dc mysql replication was a standard thing but I haven't run it myself. I guess the root of the problem is wordpress being a pig, and there's not much to do about that.
If you've got 2 racks with replication in the same DC on separate power and switches, that's probably better than the usual 1-server setup.
Right, that would likely be my target if I did an HA product.
To be honest I'm not sure if I'd do HA on KVM or use something like Hyper-V for it. KVM's HA is just so janky.
I'm still digging through litespeed logs. There's no coredump's so it isn't a hard crash. wswatch isn't reporting very many restarts from crashes either (most nodes have had a single one in 2018) so I'm kinda stuck at this point.
Francisco
Usually you don't get coredumps unless you configure some ulimit settings to allow them. Does litespeed come with any sourcecode at all? Maybe you could run a local script to check the socket status once a second (/proc/net/whatever, look at the netstat man page) and log any weirdness. That might tell you when stuff is going wrong.
The enterprise builds are completely closed off.
Francisco
What's so great about litespeed anyway? I don't know any other sites that use it, including some pretty big ones.
I don't know of anyone running it stand alone but on shared it makes a fairly big difference in performance.
Francisco
++
Even though I praised Engintron at first, LiteSpeed has been much better for me. Definitely makes a difference.
We tried Engintron and while it was an interesting idea Litespeed has done laps around it in the end.
Francisco
Uses less CPU and RAM. Comes with a better than nothing ddos protection and wordpress bruteforce protection.
Comes with quic and with http2 faster than apache's.
If you add lscache plugin to wordpress, its a great caching plugin with off server image optimization, reducing server load on shared.
Thats all I can think off, but without it francisco would only be able to get about 50% of the users he has right now.
Even when we have nodes that are running far less than half what a full node does, Litespeed still makes a great difference.
Also, at this point people look for it on their check list when shopping for shared/reseller hosting. You're paying not only for the software but for the marketing you can do because of it.
Francisco
My VPS and shared hosting went down just before 5:00 pm Pacific and just came back online a few minutes ago. Also noticed that Byvm.net was also offline. Anyone else affected?
from 21:58:29 to 22:15:43 UTC -4
It is still down for me currently. It came back shortly, but it’s been down for close to an hour.
No need for LiteSpeed. Engintron is enough.
Actually Apache with mpm fork (PHP 7.1+) or php fpm works just fine. Managed and still do manage a bunch of very high traffic servers with that setup. I offered to help you, but you didn't want my help.
Doesnt Engintron have issues with SSL certificates, where it requires a manual nginx reboot?
They have a monitor script in place that auto restarts things.
I had a lot of issues with cloudflare.
Francisco
You should know better than I. I did test it but I only use Apache and php-fpm. It does seem that they need to polish it before its actually considered production ready.
They why would Francisco risk loosing the reputation that he got, and probably a few customers due to unforeseen bugs, just to save $32/$46 per server?
Also, having the LiteSpeed team deal with bugs is way less painful than doing it yourself or creating an issue report on github and beg to have it solved asap because you are "loosing milions of dollars".
Whenever I have time Id like to test Enginetron vs LiteSpeed vs Apache, and report on any Enginetron bugs.
Engintron is still held back greatly because of Apache anyways. LiteSpeed definitely wins from my experience.
Nobody's talking about the flustercuck which is Caddy anymore?
I mentioned above my setup. It works well with no downtime. Had servers pushing thousands of TBs per month with no similar issues. Same applies to reloading the config. Even an overloaded server I once managed that was hosting 5k+ domains took seconds to restart apache. There has to be something wrong with Francisco's setup. I have also asked for the internal ticket ID so I could help escalating it to the right guys in mere minutes. I was ignored. I happen to use cPanel since 2004, that is, 14 years and am also part of their beta testers team for UI. Whenever I try to help Francisco he simply ignores.
Maybe it's time to take the hint.
As @MikeA mentinoned, litespeed is still ahead in performance, as apache continues to exist.
But as a free alternative, Enginetron is great, it's a great project and I hope one day cPanel integrates nginx officially.
If there is or not something wrong with Francisco setup, that's up to him to decide.
Nginx support is being worked on already.
From what I could read, there must be something wrong yeah.