Hosteroid downtime of 6 hours

blorged · January 2025

it really isn’t reasonable to expect or demand a single cheap machine to have high uptime, ultimately.

barbaros · January 2025

@blorged said:
it really isn’t reasonable to expect or demand a single cheap machine to have high uptime, ultimately.

Okay so let's discuss about this. If you are paying 10 bucks a month for a server I believe it's still cheap and it deserves like 1 hour downtime per week.

Do you agree or not?

blorged · January 2025

@barbaros said:
Okay so let's discuss about this. If you are paying 10 bucks a month for a server I believe it's still cheap and it deserves like 1 hour downtime per week.

Do you agree or not?

Single machine systems will fail unless you’re talking mainframe money, so yes. Spend €10 000 on a single dl360 and I’ll still have no sympathy for your downtime.

barbaros · January 2025

@blorged said:

@barbaros said:
Okay so let's discuss about this. If you are paying 10 bucks a month for a server I believe it's still cheap and it deserves like 1 hour downtime per week.

Do you agree or not?

Single machine systems will fail unless you’re talking mainframe money, so yes. Spend €10 000 on a single dl360 and I’ll still have no sympathy for your downtime.

So should I buy the whole datacenter just to host some site on production?

blorged · January 2025

@barbaros said:

@blorged said:

@barbaros said:
Okay so let's discuss about this. If you are paying 10 bucks a month for a server I believe it's still cheap and it deserves like 1 hour downtime per week.

Do you agree or not?

Single machine systems will fail unless you’re talking mainframe money, so yes. Spend €10 000 on a single dl360 and I’ll still have no sympathy for your downtime.

So should I buy the whole datacenter just to host some site on production?

No, you should accept downtime or pay for a very high availability system - that extra money would be for varying amounts of “engineering effort” (eg multiple fairly cheap VPSes with DNS load balancing health checking), “oncallers” (either to fix the system or fail over to something else) or more reliable systems (not bargain basement VMs, eg an expensive aws system with ELBs, or mainframes or whatever).

barbaros · January 2025

@blorged said:

@barbaros said:

@blorged said:

@barbaros said:
Okay so let's discuss about this. If you are paying 10 bucks a month for a server I believe it's still cheap and it deserves like 1 hour downtime per week.

Do you agree or not?

Single machine systems will fail unless you’re talking mainframe money, so yes. Spend €10 000 on a single dl360 and I’ll still have no sympathy for your downtime.

So should I buy the whole datacenter just to host some site on production?

No, you should accept downtime or pay for a very high availability system - that extra money would be for varying amounts of “engineering effort” (eg multiple fairly cheap VPSes with DNS load balancing health checking), “oncallers” (either to fix the system or fail over to something else) or more reliable systems (not bargain basement VMs, eg an expensive aws system with ELBs, or mainframes or whatever).

Just so I can host my Vaultwarden container? I am confused.

zGato · January 2025

@barbaros said:

@zGato said:

@artxs said:

@zGato said:
Let me guess, $7/yr plan?

So you expect stability and reliability from $100/month plan but shitty service from a $7/yr plan regardless or the brand? Brand doesn't matter at all to you?

For me, I got like 11 idlers from BF and Christmas deals this year, my personal record. If I like the provider I would be willing to check their more expensive offers or recommend around. But if the service is shit, I wouldn't even bother being a fully paying customer.

You're going to classify @Hosteroid's services are shit just because of this?

Also a 6 hour down time is way too long, even you pay 7 bucks per year. If a server is down more than an hour it would raise the alarm on my side. It means it's either a really serious issue (which can happen in any company like hardware or network loss etc.) or no one is bothered to keep an eye on the servers.

6 hour downtime that started at around 3AM Bucharest time is okay, specially for them because they don't advertise 24/7 technical support anywhere, nor SLA.
That doesn't mean they don't try their hardest to have the highest uptime/reliability possible. I know @Hosteroid quite well, and he's very dedicated to providing very reliable services. My VM in Bucharest from them has like 1.5yrs of uptime already.

Edit: I just filed a ticket to Hosteroid when I woke up to ask wth happened that caused for 6 hours and why no one took a look at it. I didn't scream at them, I didn't claimed I am losing million of dollars because server is done. But I believe a customer should be informed if there is long time downtimes (which Hosteroid did later).

Just by opening a ticket demanding explanations proves your shitty attitude.

Levi · January 2025

What’s more interesting is how single client could take down whole node by exploiting smth in the kernel.

naphtha · January 2025

@ralf said:
Yeah, abusing CPU resources (by which I'm assuming they mean running at 100% for a long time) doesn't by itself cause kernel panics. Faulty hardware generally causes kernel panics, it's just they're more likely when the system is under load.

If one user can do this, then so can 50 users doing normal, non abusive things. Hopefully the root cause was actually identified and fixed...

there are rare linux kernel vulnerabilities that can cause this, e.g.
https://forum.proxmox.com/threads/proxmox-virtual-environment-security-advisories.149331/#post-688609

PSA-2024-00008-1: kernel: DoS via short ethernet frames over tun/tap interfaces

The tun and tap network drivers in the Linux kernel lacked verification for short frames in XDP (eXpress Data Path).

This could be abused by a malicious guest with a VirtIO-net device to cause out-of-bound access in the host kernel and with certain hardware, even cause a kernel panic.

an out of bound access in the host kernel can cause much worse things than just a kernel panic which is why i don't understand why some customers are so obsessed with long uptimes, a long vm uptime usually means a hypervisor kernel that hasn't been upgraded in a looooong time (unless they live migrate vms to other nodes or live patch the kernel which also isn't perfect and has certain limitations)

barbaros · January 2025

@zGato said:

@barbaros said:

@zGato said:

@artxs said:

@zGato said:
Let me guess, $7/yr plan?

So you expect stability and reliability from $100/month plan but shitty service from a $7/yr plan regardless or the brand? Brand doesn't matter at all to you?

For me, I got like 11 idlers from BF and Christmas deals this year, my personal record. If I like the provider I would be willing to check their more expensive offers or recommend around. But if the service is shit, I wouldn't even bother being a fully paying customer.

You're going to classify @Hosteroid's services are shit just because of this?

No, I am happy with my idle server with Hosteroid so far. I didn't mention any single provider when I said that (and I have like some cheap servers from 7-8 providers from LET)

Also a 6 hour down time is way too long, even you pay 7 bucks per year. If a server is down more than an hour it would raise the alarm on my side. It means it's either a really serious issue (which can happen in any company like hardware or network loss etc.) or no one is bothered to keep an eye on the servers.

6 hour downtime that started at around 3AM Bucharest time is okay, specially for them because they don't advertise 24/7 technical support anywhere, nor SLA.
That doesn't mean they don't try their hardest to have the highest uptime/reliability possible. I know @Hosteroid quite well, and he's very dedicated to providing very reliable services. My VM in Bucharest from them has like 1.5yrs of uptime already.

Edit: I just filed a ticket to Hosteroid when I woke up to ask wth happened that caused for 6 hours and why no one took a look at it. I didn't scream at them, I didn't claimed I am losing million of dollars because server is done. But I believe a customer should be informed if there is long time downtimes (which Hosteroid did later).

Just by opening a ticket demanding explanations proves your shitty attitude.

I believe you never had any shitty attitude in your whole life. Asking why server was down for 6 hours isn't even a shitty attitude.

If you are a Hosting Provider, just state your provider so I won't buy any services from you. Because I mean by accident you would mark my ticket as a shitty attitude.

cybertech · January 2025

@DeusVult said:
6 hours is ridiculously long.wow!

have you used Virmach?

Arirang · January 2025

@barbaros said:
If you are a Hosting Provider, just state your provider so I won't buy any services from you. Because I mean by accident you would mark my ticket as a shitty attitude.

If I was a provider, I also would not accept your order. In future Before ordering, please add your expectation( I need full detail on anydowntime even weekend) and refer to this article in additional note section.

kmm996 · January 2025

@barbaros said:

@zGato said:

@artxs said:

@zGato said:
Let me guess, $7/yr plan?

So you expect stability and reliability from $100/month plan but shitty service from a $7/yr plan regardless or the brand? Brand doesn't matter at all to you?

That's not my point, but as how $7/yr LET users are the most annoying regarding "my vps is down dude you need to fix it right now!"
I would expect someone paying $100/m to complain about downtime, even for 5 minutes, but OP's attitude towards Hosteroid is quite shit.

Such client attitudes are the reason we're getting less $7/yr deals from providers.

Sorry I don't agree. I get your point. But as this is VPS, down server would most likely mean a node issue or neighbors abusing the server etc. which has impact on just more than one customer. If other customers paid 7 bucks or if they are expensive customers or not, that's up to the hosting company.

If an issue causes more than handful of customers than it should be treated seriously, what they paid is not related. A customer that paid 100 months per month and a customer that paid 7 bucks per year has same commenting power. They can both go to Trust Pilot, LET, LES and complain about the brand, which causes a damage.

For me, I got like 11 idlers from BF and Christmas deals this year, my personal record. If I like the provider I would be willing to check their more expensive offers or recommend around. But if the service is shit, I wouldn't even bother being a fully paying customer.

Also a 6 hour down time is way too long, even you pay 7 bucks per year. If a server is down more than an hour it would raise the alarm on my side. It means it's either a really serious issue (which can happen in any company like hardware or network loss etc.) or no one is bothered to keep an eye on the servers.

And the data you host on the server doesn't have to be production or bring income. You can simply host a Vaultwarden instance, Or a Pi-hole, or a DNS server for your own use. Downtime of this server would cause inconvenience during my daily use.

If you still say that then don't host anything that you might need daily, then I would ask wth am I supposed to host on the servers if I can't rely on the server? Apart from idling ofc.

Edit: I just filed a ticket to Hosteroid when I woke up to ask wth happened that caused for 6 hours and why no one took a look at it. I didn't scream at them, I didn't claimed I am losing million of dollars because server is done. But I believe a customer should be informed if there is long time downtimes (which Hosteroid did later).

Overused CPU core causing kernel panic getting fixed 6 hours later is still something I don't buy but I won't stir the pot. Will just note the incident.

Last time my hosthatch server was offline for more than 18 hours, and there was no news. I tried to send a ticket and talk about it on LET, and everyone told me that the promotion server support is limited. But is it reasonable that there is no announcement for more than 18 hours of offline?
There are always people who like to make fun of this with the low price? It's just that they are not the parties involved, and they don't know how anxious when your server offline more than 10 hours.

I think cheap services can use low-end outdated hardware and have limited service support, but that at least needs to ensure that the server is normal, right?

cybertech · January 2025

[@kmm996 said](/
Last time my hosthatch server was offline for more than 18 hours, and there was no news. I tried to send a ticket and talk about it on LET, and everyone told me that the promotion server support is limited. But is it reasonable that there is no announcement for more than 18 hours of offline?
There are always people who like to make fun of this with the low price? It's just that they are not the parties involved, and they don't know how anxious when your server offline more than 10 hours.

limited promotion server support is claimed by HH not the customers. it is not reasonable for 18 hours but this is HH. if you arent happy about it, the best thing you can do is not buy services from them. i have service that will expire and wont be renewed.

but if you are talking about $7/year VPS then i would say dont expect it to be rock stable, so anything that does better is a bonus.

imo regardless of price, Greencloudvps has the best support levels. so if that's important, go with them.

barbaros · January 2025

@Arirang said:

@barbaros said:
If you are a Hosting Provider, just state your provider so I won't buy any services from you. Because I mean by accident you would mark my ticket as a shitty attitude.

If I was a provider, I also would not accept your order. In future Before ordering, please add your expectation( I need full detail on anydowntime even weekend) and refer to this article in additional note section.

I am glad you are not one and hope that never get to be one.

barbaros · January 2025

@cybertech said:

[@kmm996 said](/
Last time my hosthatch server was offline for more than 18 hours, and there was no news. I tried to send a ticket and talk about it on LET, and everyone told me that the promotion server support is limited. But is it reasonable that there is no announcement for more than 18 hours of offline?
There are always people who like to make fun of this with the low price? It's just that they are not the parties involved, and they don't know how anxious when your server offline more than 10 hours.

limited promotion server support is claimed by HH not the customers. it is not reasonable for 18 hours but this is HH. if you arent happy about it, the best thing you can do is not buy services from them. i have service that will expire and wont be renewed.

but if you are talking about $7/year VPS then i would say dont expect it to be rock stable, so anything that does better is a bonus.

imo regardless of price, Greencloudvps has the best support levels. so if that's important, go with them.

I mean, we are changing the subject here but where do we draw the line?

$7 / year -> down time is expected

$10 / year -> is down time expected?

$12 / year -> is this VPS allowed to have down times?

You can find non flash deals, 1 CPU 1 RAM 10 GB disk for 12 bucks per year on LET. If you stick to US they are even more cheaper. Are those considered promotion? Are these packages allowed to have downtimes?

I mean after today, if someone says that they pay 10 bucks per month for their service and they complain about downtime, I would say it's expected for that cheap servers to have downtime. My new standard is a server should cost 50 bucks per month for a downtime to not be expected. Cheaper than that is considered cheap ass service.

There are "special" customers on LET, after today some consider I am one also, that put their all business on cheap ass servers and complain that they lose money. Like I said first line of my message on LET, the server was an idler. But in any case, a client has rights to ask cause of downtime. IF they pay 5 bucks per month or year. I am not saying a client should expect 99.99% uptime on cheap ass servers but after all a customer pays money to get such service.

Or are we at this point that just because you bought some cheap deal (flash sales or BF deals etc) you should just STFU and accept what's coming your way?

Edit: To add, if it's not OK to ask your provider why the server is down for 6 hours, then I believe it's not OK to ask your provider if a neighbour is abusing the CPU and you can't even use your server, after all it was a cheap ass server.

On top of that, I believe when customers complain in provider threads that their server is too slow (like storage server deal providers) instead of providers trying to find a solution, they should tell their customer to STFU and be happy that they got a chance to buy this cheap server. I am talking about 7 euro per year 1 TB storage, or like 20 euro per year storage servers.

kmm996 · January 2025

@cybertech said:

[@kmm996 said](/
Last time my hosthatch server was offline for more than 18 hours, and there was no news. I tried to send a ticket and talk about it on LET, and everyone told me that the promotion server support is limited. But is it reasonable that there is no announcement for more than 18 hours of offline?
There are always people who like to make fun of this with the low price? It's just that they are not the parties involved, and they don't know how anxious when your server offline more than 10 hours.

limited promotion server support is claimed by HH not the customers. it is not reasonable for 18 hours but this is HH. if you arent happy about it, the best thing you can do is not buy services from them. i have service that will expire and wont be renewed.

but if you are talking about $7/year VPS then i would say dont expect it to be rock stable, so anything that does better is a bonus.

imo regardless of price, Greencloudvps has the best support levels. so if that's important, go with them.

I understand limited service and I accept it. I just don't understand some providers prefer to remain silent. They could post an announcement or set up a network status page to let users know what's going on, which would be simple, efficient and responsible. I believe this would also reduce the tickets.

cybertech · January 2025

@barbaros not diving into such deep details but i hope you get the gist of "$7/yr" .

there are many providers claiming 24/7/365 support including HH, but reality always hits differently.

many of LE providers are small businesses, and knowing who the one-man-operations are will help buyers manage their expectations.

i like hosteroid for their M247 in some locations, and AMS is new for them, i would personally expect teething issues.

would i continue to buy their promos? hell yeah.
but would i put production on them? definitely not yet at least.

barbaros · January 2025

@cybertech said:
@barbaros not diving into such deep details but i hope you get the gist of "$7/yr" .

Sure thing. I just expect a downtime announcement when it happens for a long time. Not a personalized one. A provider here (I won't give out names as it's not relevant) sent an email to their clients saying that there is an ongoing issue with their services. I believe this should be the standard. Or even personal users use services like Uptime Kuma, which is a centralized way to show ongoing issues.

there are many providers claiming 24/7/365 support including HH, but reality always hits differently.

It's not even really possible to expect 24/7/365 support without paying premium in big providers. Yet, just to point that out, I filed 2 tickets to Hosteroid (one was my fault because I forgot to choose OS) and they were really fast to reply (meaning 1 hour or less response time).

would i continue to buy their promos? hell yeah.
but would i put production on them? definitely not yet at least.

that I do agree also, and to add, I would say the same thing for any provider with such deals, this wouldn't be specific to Hosteroid.

admax · January 2025

@barbaros said:
Should it really take 6 hours to fix this issue, I’m not sure, I will just keep monitoring my idle server.

Effectively Avoid Missing Popcorn Events

suyadi92 · January 2025

@cmeerw said:
Luckily it was only a kernel panic and not a takeover of the host machine...

No it's just a kernel PTSD but takes much longer to recover

truemagic · January 2025

I like this kind of post to reassure me which provider I should go with for a "production" environment, regardless how much I pay for it.

CloudHopper · January 2025

To be fair to the OP, I think if a host has a 6 hour outage then they should communicate it to their customers...and I'm paying Hosteroid €21/year, (~$25/year), so it's hardly a $7/year service.

On the other hand, my server with them in Vilnius has been rock solid and I wouldn't worry about using it for Production because this is the first complaint I've heard about their service.

For me, Hosteroid is offering excellent value and decent quality. I won't be renewing annual deals with a few of the LET 'big name' providers, but this one is definitely a keeper.

donli · January 2025

@CloudHopper said:
To be fair to the OP, I think if a host has a 6 hour outage then they should communicate it to their customers...and I'm paying Hosteroid €21/year, (~$25/year), so it's hardly a $7/year service.

€21/year for what specs?

CloudHopper · January 2025

@donli said:

@CloudHopper said:
To be fair to the OP, I think if a host has a 6 hour outage then they should communicate it to their customers...and I'm paying Hosteroid €21/year, (~$25/year), so it's hardly a $7/year service.

€21/year for what specs?

This offer here. It's been rock solid
https://lowendtalk.com/discussion/195070/hosteroid-vilnius-offer-giveaway-50gb-nvme-4-gb-ram-21-42-year/p1

jaden · January 2025

HetrixTools reported another hour of downtime at 9pm UTC today (12 Jan). Curious to know if it was due to another user abusing the CPU.

barbaros · January 2025

@jaden said:
HetrixTools reported another hour of downtime at 9pm UTC today (12 Jan). Curious to know if it was due to another user abusing the CPU.

Not sure if we are on the same nodes or if you got the email but there was planned maintenance to fix this issue. I just think that downtime was related with that. So all should be good.

barbaros · January 2025

I will give an update to this matter. Apparently Hosteroid is still having issues with Amsterdam node caused by their colocation provider (their saying). Even after the maintenance, server still got down for an hour during night time (GMT).

To add, they are quickly responding, and offered a suggestion to move my VPS to another location in Bucharest, Bratislava, Vilnius. So if you are planning to order from their Amsterdam location, keep this in mind (for VPS).

ralf · January 2025

@barbaros said:
I will give an update to this matter. Apparently Hosteroid is still having issues with Amsterdam node caused by their colocation provider (their saying). Even after the maintenance, server still got down for an hour during night time (GMT).

Interesting, my AMS Hosteroid seems to have been back up for a bit over 19 hours at the moment. I does seem to have had a lot of unclean shutdowns since I got it.

# last |grep reboot
reboot   system boot  6.1.0-9-amd64    Sat Jan 25 06:54   still running
reboot   system boot  6.1.0-9-amd64    Fri Jan 24 21:43   still running
reboot   system boot  6.1.0-9-amd64    Thu Jan 23 11:44   still running
reboot   system boot  6.1.0-9-amd64    Sat Jan 18 13:27   still running
reboot   system boot  6.1.0-9-amd64    Sat Jan 18 13:21   still running
reboot   system boot  6.1.0-9-amd64    Sun Jan 12 11:13   still running
reboot   system boot  6.1.0-9-amd64    Sun Jan 12 07:37   still running
reboot   system boot  6.1.0-9-amd64    Sat Jan 11 09:23   still running
reboot   system boot  6.1.0-9-amd64    Thu Dec 26 02:09   still running
reboot   system boot  6.1.0-9-amd64    Mon Dec 23 10:59   still running

To add, they are quickly responding, and offered a suggestion to move my VPS to another location in Bucharest, Bratislava, Vilnius. So if you are planning to order from their Amsterdam location, keep this in mind (for VPS).

My Bucharest one was up for 4 days and before that 20 days. All were clean shutdowns (although the last wasn't by me!)

Blembim · January 2025

I wish NL would be more stable from now on.
Given that their upstream network is prem, it would be sad if i could't use it for anything cause of frequent downtime

Howdy, Stranger!

Categories

In this Discussion

Hosteroid downtime of 6 hours

Comments

Howdy, Stranger!

Quick Links

Categories

In this Discussion

Hosteroid downtime of 6 hours

Comments