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
and what about lv-shared01?
Every shared in LV had similar stuff with the two I mentioned above.
The 1 minute blips are still happening, I got one on my monitors in the past day or so. There is some users with a lot of domains, they've been ticketed to see if we can cut that back some. There's still no confirmation that it's related to that, but they shouldn't be running those services on shared anyway.
Francisco
You're talking about @Francisco revenue per server, I thought we were talking about changes that positively affect his clients, not his pocket.
My point is, overselling is a bitch, and if by limiting the number of domains per account results in more accounts per server then I'm guessing clients won't see much of an Improvement if any.
I don't want to discuss @Francisco business plan, not my place, I know nothing about it, and honestly it seems to me that he is doing quite well, I've seen hosting companies not coping with a few servers renting bills, while @Francisco is about to dish out 40k for HDs alone.
I understand your point of view, I really do, but not sure if it's fair to expect @Francisco to come up with a service on par with other companies that charge several times over what he does.
If you want, I can refer you to a company that rarely haves any downtime (but they still have the occasional downtime, reality sucks). But you need to prepare yourself to pay several times over what you're paying Francisco.
To be fair, I don't think what @suriam is asking is unreasonable: inexpensive reliability for low usage, non-abusing clients.
BuyShared is strongly recommended on this forum, and @Francisco's responses and responsiveness here show why. Perhaps the expectation built up by this sets people up for disappointment when seeing frequent interruptions, even if they are not major outages.
On lu-shared04, since 2017-12-06 I've had UptimeRobot report 12 outages ranging from 1 to 38 minutes. In comparison, with another host that doesn't get discussed as often but has the same pricing, since 2017-11-29 I've seen only 2 interruptions for 2 and 37 minutes.
Right, but I've already discussed that lv-shared04 doesn't have the issue with 'excessive domains' as a couple resellers do, which still leads me to believe it's something else acting up.
-15|48|14- :: lv-shared04 ⋟ # uptime 15:48:14 up 10 days, 10:59, 2 users, load average: 10.40, 10.81, 11.79 -15|48|35- :: lv-shared04 ⋟ # free -m total used free shared buff/cache available Mem: 72314 27990 1139 5209 43184 38257 Swap: 0 0 0The node has plenty of CPU & RAM spare. There's no swap, but that's not needed for a node with 30 - 40GB of RAM spare. Swappiness=0 is also set.
lu-shared04 is pretty similar, plenty of RAM and a good bit of spare CPU.
Francisco
@Francisco glad to see you're addressing issues. The details are interesting to the LET crowd, but I think the average user just cares if their site is up and works.
Tbh, considering BuyShared's reputation, the number of interruptions surprised me. I should also mention there have been zero issues so far in 2018. No problems noticed with load or ram.
There was a spike earlier on lu-shared04 but that ended up being someone bruting the ever living hell out of cPanel itself. They're still hammering away but they're blocked.
The blips are out of our control. It isn't like the box is rebooting every time or something like that, you can check
uptimeand see for yourself. There's no OOM's for node side things (but plenty for end users that hammer their plans too hard), so it's not that. I've checked.I dunno, it bugs the hell outta me and has been something I've spent a good bit of time trying to track down. Every time we've seen it get fixed, there's a patch from either cPanel or Litespeed that relates to it.
Francisco
Cool, how? Do you allow shell access?
Absolutely, always have. You can use port
2222.Francisco
Any chance of the 1G plans coming back in stock soon?
You can open a ticket and Karen will check her stock.
Things are pretty sold out right now, I think we got a tiny bit of stock on slices in LV and some LU resellers, but that's it.
Francisco
The reputation comes from @Francisco posting detailed RFOs and IDKs to this forum. This is mostly a forum of sysadmins and so everybody appreciates the admin'ing process done competently and transparently. . .
. . . but People you are missing the essence here: Buy* gets more complaints on this forum than others, and @Francisco comes back every single time and shows 110% customer service, and that is what defuses the original complaints every single time.
Buy* is a great service for what you pay, but you just don't pay much to start, and there is an economizing process that makes these prices viable which includes overcommitment, cost-cutting on support, not patching for the Meltdown + Spectre issues, etc. . .
Im interested in the Cpanel and the easy back up restore that it offers. I'll be using wordpress + caching to html files + Cloudflare caching, so the load on the server will be almost nothing. Just need it for the cpanel.
Good idea? Would these downtimes affect the cloudflare caching if files already cached by them?
BIOS updates still haven't rolled for some of our stuff, it'll be dealt with when ready. I'm not going to rush and start ripping down gear without preparing for things. Some of our stuff has kernelcare and will be patched whenever that's distributed by CloudLinux. Why haven't you hated on every provider out there using KernelCare that hasn't done reboots?
People do complain about us because they felt they would get something out of it. Many providers will rush to throw credits and refunds and all the likes to try to dodge from this. Many of these providers are long since gone yet here we are, continuing on as always. We do well in the polls, I give countless hand outs in code, help, commentary, and even my own spare time to fix their broken gear.
You keep trying to be smug but all you end up being is a dumbass that keeps getting banned from places in one motion, and then begging for a job in another. Do we really need to go through the mountains of logs from #frantech where you sat going on about how I need to hire you instead of random worker you don't like? Or how you insisted on emailing me said resume?
I'm sure i'm not the only provider.
Francisco
There as 2 bigger ones that happened in the past week or so, past that we're talking things are < 1 minute from Litespeed restarts. We documented what caused the problem, and have applied short term, and planned long term, fixes for it.
They're not a constant thing that happens every 5 minutes. In the past 3 days I've had 2 I think, with 1 of them being me (hopefully) getting coredumps enabled.
We have many thousands of shared customers, we wouldn't have those if things were melting a hole in the ground. We continue to improve thing and in a few cases have been able to point out bigger edge cases that providers will run into.
Any provider that thinks they might be seeing the same problem and thinks they're in my good books is welcome to contact me and I may share my detection scripts.
There's a 3 day refund policy for you to try things out. This doesn't apply to crypto payments for obvious reasons.
Francisco
Do you accept any crypto currencies other than Bitcoin?
Yep, we accept 2 dozen or so coins. All the major alt's are supported.
Francisco
Dint mean to imply that this happens all the time. Just wanted a worse case scenario check and if the measures above will survive any such downtime, even if once a year or decade. Also wanted suggestions on whether adding more redundancy is possible.
If you're all static content then you could do something with CF, sure.
Anyway, thanks for the interest!
Francisco
@jiggawattz Did you feel the burn? I sure did!
@Francisco wants to divert attention from brand-negativity toward petty IRC drama, and I'm not sure how to take being called a dumbass, but I'll address the former and ignore the latter as just a very frustrated @Francisco
Nearly every provider uses KernelCare, and indeed all my other providers rebooted. I'm not even sure if KernelCare will be able to release a reboot-less patch.
Nevertheless, your stuff is vulnerable. Let's not get distracted. Other providers toiled over this.
This is true in some situations, but the outages described in this thread are real, and you have some pretty loyal customers imho. As I have said before, you've earned that by being transparent on LET.
You posted a job advertisement on WHT and I responded with my CV. I'm not sure who is "begging" here.
I offered my help on IRC, pitching myself as someone 1) who would probably cause less negative LET threads than @Aldryic did and 2) available during EU working hours when you sleep.
Mind you, I rescind my offer to help you. You can find someone else.
You are indeed the only provider I use who hasn't applied the Meltdown + Spectre patches. I respect your reasoning, but it's note-worthy.
Not really - because most of my VPS are patched against the Meltdown. (Except the VPS on BuyVM, of course.)
1 - KernelCare stuff didn't work without a reboot anyway;
2 - It's too soon to upgrade unless microcode is all ready, etc;
3 - No, not all Providers have updated their systems. Some have, and still aren't protected for the 3 different attacks;
4 - @Francisco would hire me instead.
Uh huh... Because uptime is measured in less than a week periods...
@sureiam Are you paying $5 / year ?
The total for the account is much higher than $5 a year as I've added multiple accounts for end users vs just putting them all under one account. Unfortunately that didn't protect vs downtime as they all ended up on the same node. However at this stage I'm thinking a VPS would be much more reliable.
You would definitly have more direct control however keep in mind the licenses you'd need to pay with a vps though.
VPS + cPanel + JetBackup + CloudLinux + Softaculous + KernelCare(?)
Softwares themselves will cost him above 35$ then cost for Good VPS separately
No need for CloudLinux if stuff's done right. :>
Naa. I have solutions for those that won't cost much. SFTP would be enough for their needs. They don't needd cpanel or softaculous as i previously assumed. Cloudlinux and Kernalcare are also unnecessary with proper configuration and management. Only if your dealing with very large number of users which I'm not. Backup also isn't an issue. I can get it done for about the same cost but now i have to manage it which i didn't want to do. However it seems the alternative isn't as dependable as I'd hoped.