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
Many people sell hosting as side business to customers . Specially they do web development as primary work and so customers buy whatever plan they offer , so when any work needed , they will be available always . These guys usually sell reseller this kind of hosting .
Nah, the other node dying earlier this year was longer. Well, if you had to wait for a restore it took awhile. Shit happens.
I agree!
I allow numerous Addon domains which are helpful for small website designers. I offer them support limited to what I can do in WHM/cpanel. The rest, I will open a ticket and wait (like yesterday's outage). Netcetera, Onra, Frantech helps me in doing such small hosting business. Hope you checked my whois and found it as 2006 registered. No special marketing.
Hostdare said it right. Thank you!
Pretty sure I've seen this mentioned a bunch on other threads.
Francisco, you could run two PHP daemons, one for long-standing domains and an 'edge' one with a few dozen, and migrate the edge on to long-standing every week or so. It'd affect less domains. There's probably a better way, just thinking aloud.
Actually they had fixed this bug a few versions back, but it seems it has returned. I think it's more a case of cPanel and Litespeed being out of sync again.
Last time cPanel had to give me a patch as a hold over since their whole graceful system just checks the availability of certain files being in specific places (namely /usr/local/lsws having a bunch of stuff in it). The problem is that, at least in the patch they gave me, they were keeping a static list of all the libraries they were trying to detect.
This isn't a huge issue if you got a small amount of accounts or domains, but it isn't that uncommon to find a user with a few hundred sub domains and such under a reseller. Do that a few times and your httpd.conf is going to get very fat, very quickly.
Francisco
This applies only to the free version. The paid version checks every 1 min. I turned UptimeRobot off for BuyVM. Too many downtime notifications.
I really got bad view of litespeed now. There have been many downtimes ar buyshared that comes from litespeed. Is litespeed really that bad? Have only buyshared this issue or all providers using it?
Netcetera too uses Litespeed. But thats much much better in case of uptime.
Thanks for your great feedback
Have never experienced this at all. But as Francisco said, it sounds like it's happening when there are a shitload of sites on the server.
We never had an issue with LiteSpeed - might depend on the scale/number of domains per server though.
Why does anyone still use shared hosting? You can probably get a VPS for only a few pennies more that is probably 10x better.
because
Most of those people probably don't want to deal with the overhead of managing a server and just want some one click type panel directly installed.
You can get fully managed VPS with one click install and config stuff as well. I don't think it's any harder to manage.
cPanel licence have cost. not everyone know server management. and many even don't know how to install free panel like vestacp.
Cpanel one month license cost+vps management = one year shared hosting, free support, free SSL, cloudlinux etc
Shared Hosting is usually much cheaper than cPanel license itself.
Litespeed is fairly solid, we just push is really hard so we likely see some bugs come up that most don't. The hard reboot thing is a cPanel issue since they're still working through their integration with it and sometimes run into issues. Some of the biggest issues we had with it required auto restarters to help until they got a full patch in place.
Sometimes cPanel generates a bad configuration just because there's a chance 2 processes are trying to modify the httpd.conf. This ends up with a pretty fucked httpd.conf. Litespeed is very forgiving and will usually do its best to just ignore the bad entries.
Since cPanel doesn't know the configuration is broken and Litespeed doesn't (yet) fire off a regenerate job on its own, you can have a pretty big mess until another update gets run (be it another domain action or a manual invoking). We've had cases where users sites don't load at all, or their cagefs limits aren't being enforced, or their PHP selector settings aren't being obeyed.
We did try some of the NGINX plugins and they were decent. They had some issues with cloudflare though which put that project to rest though.
In the end, I like it a lot, if I didn't I wouldn't be paying as much as we do for it.
Francisco
Not everyone has sysadmin skills.
Heck, even people like me who do sometimes still don't want the hassle of maintaining and patching a web server, managing mail, etc.
Did you open a ticket with cPanel as well? I've asked for your ticket ID/ID's so I could get those escalated / investigated further for you.
I've talked to them and they've given me patches. I think it's mostly just a cat/mouse game sometimes with their two teams and then you got us pushing it hard so the problem becomes a lot more obvious.
The restart thing has been a lot better in the past couple weeks so it's likely the latest versions have helped a lot. Before it was a half dozen times per node, per day.
Francisco
Yeah later on I noticed that my site went down after I replied saying it wasnt down
It just goes down for 60 second increments
It should reload the configuration, not restart apache. So that doesnt make much sense.
Apache requires a graceful restart for every configuration change, you can see the log of each request and how long it took to process at /usr/local/cpanel/logs/safeapacherestart_log
What was happening at the time was Litespeed wasn't processing it properly (or was ignoring the rehashes since i'm assuming it just using an inotify or something similar to monitor the file), and cPanel was forcing restarting.
Francisco
Always -HUP or -USR1.. unless you're systemd, then you're just fucked.