We will halt all sales for the next 24 hours due to the fact that the 3 newly published /48 ( that get divided into /64 for you ) are only partially visible on global BGP
Plus we have an issue with the old 2A0D:8140::/48 that our ISP accidentally removed from the filters rather then adding the new ones. - this is a more pressing issue
We will be back tomorrow, but first we need to fix our stuff
@host_c said:
We will halt all sales for the next 24 hours due to the fact that the 3 newly published /48 ( that get divided into /64 for you ) are only partially visible on global BGP
Plus we have an issue with the old 2A0D:8140::/48 that our ISP accidentally removed from the filters rather then adding the new ones. - this is a more pressing issue
We support providers that treat IPv6 as an essential service rather than secondary to IPv4.
It is honorable to halt sales when IPv6 isn't working.
We accept Karma donations for the last flan. 🍮 affbrr
@host_c said: Plus we have an issue with the old 2A0D:8140::/48 that our ISP accidentally removed from the filters rather then adding the new ones. - this is a more pressing issue
What's the current status of this more pressing issue? :P
While this is a bit off-topic, I have a question to those who know more about IPv6 than I do. I could reach the server from outside (via ssh and ICMP ping) over IPv6, however from inside the VPS I couldn't reach any IPv6.
The output from ip a was the following:
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether XX:XX:XX:XX:XX:XX brd ff:ff:ff:ff:ff:ff
inet XXX.XXX.XXX.XXX/24 brd XXX.XXX.XXX.XXX.255 scope global eth0 // allocated IPv4
valid_lft forever preferred_lft forever
inet6 2a0d:XXXX:XXXX:XXXX::XX/64 scope global // allocated IPv6
valid_lft forever preferred_lft forever
inet6 fe80::XXXX:XXXX:XXXX:XXXX/64 scope link // link-local address
valid_lft forever preferred_lft forever
inet6 2a0d:XXXX:0:4b:XXXX:XXXX:XXXX:XXXX/64 scope global dynamic mngtmpaddr noprefixroute
valid_lft XXXXXXXsec preferred_lft XXXXXXsec
inet6 2a0d:XXXX:0:125:XXXX:XXXX:XXXX:XXXX/64 scope global dynamic mngtmpaddr noprefixroute
valid_lft XXXXXXXsec preferred_lft XXXXXXsec
I have no idea where the last two addresses (on some of my VPSes from @host_c there were even more) come from, I have set net.ipv6.conf.all.autoconf, net.ipv6.conf.all.accept_ra and net.ipv6.conf.all.use_tempaddr all to 0 and DHCP was disabled.
When I deleted the two addresses with ip a del ADDRESS dev eth0, IPv6 worked from inside the VPS, but after a few seconds those addresses reappeared.
But after some time (I think around 11 PM GMT+1), the addresses weren't added anymore and IPv6 works since then.
(This is no criticism in any way, I'm just interested in how that happened.)
Now my question is: where did these addresses come from and could I have disabled that?
If this is too off-topic, mods can maybe move this post to the cest pit.
Since you push IPv6 automagically through Router Advertisements for most(?) customers, why not just simply provide the "new IPv6 settings" in the Control Panel (under "Assigned IPs"), for the few customers clever enough to manage IPv6 manually?
@Shot² said:
Since you push IPv6 automagically through Router Advertisements for most(?) customers, why not just simply provide the "new IPv6 settings" in the Control Panel (under "Assigned IPs"), for the few customers clever enough to manage IPv6 manually?
these will be available there.
but we have to update ~500+ services, I doubt 50 will set the up manually.
@lukast__ said: While this is a bit off-topic, I have a question to those who know more about IPv6 than I do. I could reach the server from outside (via ssh and ICMP ping) over IPv6, however from inside the VPS I couldn't reach any IPv6.
Is this fixed? I've been there as well. They resolved it by manually whitelisting my subnet.
@lukast__ said: While this is a bit off-topic, I have a question to those who know more about IPv6 than I do. I could reach the server from outside (via ssh and ICMP ping) over IPv6, however from inside the VPS I couldn't reach any IPv6.
Is this fixed? I've been there as well. They resolved it by manually whitelisting my subnet.
For me it works since around 11 PM GMT+1 yesterday, but I didn't need to open a ticket.
For us IPv6 nerds (who manage IPv6 manually): should we power-off the server before or after opening a ticket? it's not clear in the mail.
Is there a time limit? What happens if we don't open a ticket by then? Service cancelled?
@Shot² said:
For us IPv6 nerds (who manage IPv6 manually): should we power-off the server before or after opening a ticket? it's not clear in the mail.
Is there a time limit? What happens if we don't open a ticket by then? Service cancelled?
Nerds receive a ticket from us with the settings to set up ( "nerds" are identified by having custom install, 99% of the cases this is true, and most of them opened a ticket already )
All usual folks, we will set the new ip in cloud-init and issue a power-off / power on - they need to do nothing.
@localhost said: @host_c thanks for the /64
Any quick quides to enable ipv6 on AlmaLinux 9?
Seems when I installed via iso only ipv4 is enabled by default
@localhost said: @host_c thanks for the /64
Any quick quides to enable ipv6 on AlmaLinux 9?
Seems when I installed via iso only ipv4 is enabled by default
Thanks
nmtui and off you go.
Is there a way to use nmtui which would setup ipv6 like a dhcp does for ipv4? For me it is asking me to enter the details manually. Thnx
@localhost said: @host_c thanks for the /64
Any quick quides to enable ipv6 on AlmaLinux 9?
Seems when I installed via iso only ipv4 is enabled by default
Thanks
nmtui and off you go.
Is there a way to use nmtui which would setup ipv6 like a dhcp does for ipv4? For me it is asking me to enter the details manually. Thnx
You can set IPv6 to "automatic" and see what happens... But it depends on the network whether it works or not.
@localhost said: @host_c thanks for the /64
Any quick quides to enable ipv6 on AlmaLinux 9?
Seems when I installed via iso only ipv4 is enabled by default
Thanks
nmtui and off you go.
Is there a way to use nmtui which would setup ipv6 like a dhcp does for ipv4? For me it is asking me to enter the details manually. Thnx
You can set IPv6 to "automatic" and see what happens... But it depends on the network whether it works or not.
DHCP will do nothing on our infrastructure.
There is no such service enabled.
Ip addresses are allocated / service by a provisioning module ( that started working as it should , hence the posibility to allocate a whole /64 rather rather then 1 /127 )
@tmntwitw said: I might be wrong, but for a single IPv6 address, the subnet mask should be /128 right?
not in our setup, this whole shit started ~1 year back.
Let me give you all an insight.
We used IP Manager 2 from ModulesGarden, at that was and is awesome for IPV4 but not for IPV6, it could not allocate a whole /64 ( from a /48 or any )
So we did a temporary work around, ( these usually bit you in the ass or in other sensitive places) by setting a /64 on each subnet we have, allocating a single IP from that /64 to customer ( as that worked in the module ) and wait out the update to IPMANAGER 3.
Update came ~2 months ago, but by that time we had give out a bunch of manually assigned /64's.
Each of these had :1 set up on the router and usually :2,3,4 for the customer - yeah, I know, lame, but it is what it is.
We ditched IP Manager 2 ~2 months ago ( that was fun as it was a ton of work so you all do not loos the already assigned ip's at the moment we selected disabled module).
And by the tests we did over the past 1 month we manage to implement the correct way to allocate a /64 from a /48 set up on each IPV4 subnet. ( all IPV4 subnets are separated at layer 2, each one has it's own broadcast domain and separate IPV6 GW + firewall stuff and other )
So when our ISP took out by mistake a /48 we used before for the old setup, I said to myself that this is as any good moment to do the implementation.
And here we are, modifying all to standard so we can ditch the " open a tiket for IPV6 /64 " and also this give us the following flexibility wise:
allocate multiple /64 IPV6 at product check-out
create IPV6 only services ( save a few bucks / year )
create IPV6 only products that can select to allocate one IPV4 address if desired by the customer at check-out
I know we are not winning loyalty points stopping your services, but we had 2 options on the table:
A. send out 500-600 E-mails so customers set up the new IPV6 setup or do a power off power on to auto apply the stings - all this when they read the email. - most probably this would taken 1 Decade
B. kill the old setup, we do all the work + some of you that have "nerd" setups - this will take ~1 week by our estimates.
I chose option B, and if you want to blame someone, blame me.
I really did not want to drag with us into 2025 an "improvisation" but rather start the year clean and focus on the more important stuff.
Comments
Chicken? I want chicken.
Chicken and rice with a hole?
Damned right I will don't make send out the phone spiders!!
The Yeti has left the building.
Weird hole right? I'm going to tell them to not make holes to my food.
I think the hole is to put sauce in?
Never. We don't do that.
Fellas,
We will halt all sales for the next 24 hours due to the fact that the 3 newly published /48 ( that get divided into /64 for you ) are only partially visible on global BGP
Plus we have an issue with the old 2A0D:8140::/48 that our ISP accidentally removed from the filters rather then adding the new ones. - this is a more pressing issue
We will be back tomorrow, but first we need to fix our stuff
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
@host_c fix my stuffs!!!!!!!!!!!!!!!!!
The Yeti has left the building.
We support providers that treat IPv6 as an essential service rather than secondary to IPv4.
It is honorable to halt sales when IPv6 isn't working.
We accept Karma donations for the last flan. 🍮 affbrr
Self rdns change would be cool.
Amadex • Hosting Forums • root.hr
I will quote @imok
Rest assured that the "Marketing Department" has enough funds to sustain these crazy deals.
If you read the beginning of of this thread, I have stated that we have other measures of income also.
With that said, we will make some more
, on IPV6 only as that does lower the expense on this.
We tried to implement NAT64, but it did not go as expected , so we dropped it and focus on more important stuff.
We managed to fix the IPV6 problems ( almost all ) since last night so sales will resume shortly.
We need a little time-off by the way, so see you shortly.
HOST-C
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
LETS GO FAT32 SPECIAL DEAL
食之无味 弃之可惜 - Too arduous to relish, too wasteful to discard.
What's the current status of this more pressing issue? :P
So, will the crazy IPV6 storage VPS deal make a comeback?
yes
if you do not have connectivity by now, open a ticket, this goes for any of you reading this.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
I have multiple vps ipv6
ping ipv6.google.com
PING ipv6.google.com(sof02s48-in-x0e.1e100.net (2a00:1450:4017:811::200e)) 56 data bytes
Do I need to open a work order as well?
What is the output of
ip a?LTstats resource monitoring system | Recommended providers: Onidel, Layer7, v6node, HostBilby, Dataforest, Gigahost | AS202986
please do,1 ticket, and we will take it from there.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
Tonight? or Now?
ticket
open a ticket so we can fix these, this is not the place.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
Thank you for those that already did
we will issue an official statement via e-mail once we sort all out.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
While this is a bit off-topic, I have a question to those who know more about IPv6 than I do. I could reach the server from outside (via ssh and ICMP ping) over IPv6, however from inside the VPS I couldn't reach any IPv6.
The output from
ip awas the following:I have no idea where the last two addresses (on some of my VPSes from @host_c there were even more) come from, I have set
net.ipv6.conf.all.autoconf,net.ipv6.conf.all.accept_raandnet.ipv6.conf.all.use_tempaddrall to 0 and DHCP was disabled.When I deleted the two addresses with
ip a del ADDRESS dev eth0, IPv6 worked from inside the VPS, but after a few seconds those addresses reappeared.But after some time (I think around 11 PM GMT+1), the addresses weren't added anymore and IPv6 works since then.
(This is no criticism in any way, I'm just interested in how that happened.)
Now my question is: where did these addresses come from and could I have disabled that?
If this is too off-topic, mods can maybe move this post to the cest pit.
LTstats resource monitoring system | Recommended providers: Onidel, Layer7, v6node, HostBilby, Dataforest, Gigahost | AS202986
The RA announcement were Meade wrongly by us. In attempt to preserve old configuration and the new one for /64 allocation
Check your e-mail, we already announced that we will modify all services to /64 from /127
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
Since you push IPv6 automagically through Router Advertisements for most(?) customers, why not just simply provide the "new IPv6 settings" in the Control Panel (under "Assigned IPs"), for the few customers clever enough to manage IPv6 manually?
Happy customer at AlexHost, AxusHost, Bakker-IT, Host-C, Ionos, Veesp + NanoKVM ftw.
When will the IPv6 network update be implemented?
these will be available there.
but we have to update ~500+ services, I doubt 50 will set the up manually.
It has already started. It will take some time .....
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
As a pain i the but this is for most of you, we wished tu finalize these ASAP, so we can all enjoy quiet holidays and other “stuff”
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
Keep up to good work man I for one appreciate it
The Yeti has left the building.
Is this fixed? I've been there as well. They resolved it by manually whitelisting my subnet.
For me it works since around 11 PM GMT+1 yesterday, but I didn't need to open a ticket.
LTstats resource monitoring system | Recommended providers: Onidel, Layer7, v6node, HostBilby, Dataforest, Gigahost | AS202986
@host_c you fixering the Yeti's crapola today?
The Yeti has left the building.
Please open a ticket, forum is NOT* a valid place to ask for a support.
/s
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
Flagging
Have no clue, I left STU to fix as much as he can. I am out of town.
He is going over services 1 by 1
I will join him tomorrow on this.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
Tell STU that you will sell him to the Yeti if he fails!!!
The Yeti has left the building.
told him.
Now, going back to tickets, only a few hundred services remaining to fix.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
For us IPv6 nerds (who manage IPv6 manually): should we power-off the server before or after opening a ticket? it's not clear in the mail.
Is there a time limit? What happens if we don't open a ticket by then? Service cancelled?
Happy customer at AlexHost, AxusHost, Bakker-IT, Host-C, Ionos, Veesp + NanoKVM ftw.
Nerds receive a ticket from us with the settings to set up
( "nerds" are identified by having custom install, 99% of the cases this is true, and most of them opened a ticket already )
All usual folks, we will set the new ip in cloud-init and issue a power-off / power on - they need to do nothing.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
@host_c thanks for the /64
Any quick quides to enable ipv6 on AlmaLinux 9?
Seems when I installed via iso only ipv4 is enabled by default
Thanks
nmtuiand off you go.dnscry.pt - Public DNSCrypt resolvers hosted by LowEnd providers • Need a free NAT LXC? -> https://microlxc.net/
Is there a way to use nmtui which would setup ipv6 like a dhcp does for ipv4? For me it is asking me to enter the details manually. Thnx
You can set IPv6 to "automatic" and see what happens... But it depends on the network whether it works or not.
dnscry.pt - Public DNSCrypt resolvers hosted by LowEnd providers • Need a free NAT LXC? -> https://microlxc.net/
DHCP will do nothing on our infrastructure.
There is no such service enabled.
Ip addresses are allocated / service by a provisioning module ( that started working as it should , hence the posibility to allocate a whole /64 rather rather then 1 /127 )
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
lets goooooooooooooooooooooooooooooooooo
I bench YABS 24/7/365 unless it's a leap year.
I am happy to report, we have a lot of 'nerds', much more then I anticipated. I am proud of you all
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
I might be wrong, but for a single IPv6 address, the subnet mask should be /128 right?
This incident should definitely be recorded in HOST-C's history.
not in our setup, this whole shit started ~1 year back.
Let me give you all an insight.
We used IP Manager 2 from ModulesGarden, at that was and is awesome for IPV4 but not for IPV6, it could not allocate a whole /64 ( from a /48 or any )
So we did a temporary work around, ( these usually bit you in the ass or in other sensitive places) by setting a /64 on each subnet we have, allocating a single IP from that /64 to customer ( as that worked in the module ) and wait out the update to IPMANAGER 3.
Update came ~2 months ago, but by that time we had give out a bunch of manually assigned /64's.
Each of these had :1 set up on the router and usually :2,3,4 for the customer - yeah, I know, lame, but it is what it is.
We ditched IP Manager 2 ~2 months ago ( that was fun as it was a ton of work so you all do not loos the already assigned ip's at the moment we selected disabled module).
And by the tests we did over the past 1 month we manage to implement the correct way to allocate a /64 from a /48 set up on each IPV4 subnet. ( all IPV4 subnets are separated at layer 2, each one has it's own broadcast domain and separate IPV6 GW + firewall stuff and other )
So when our ISP took out by mistake a /48 we used before for the old setup, I said to myself that this is as any good moment to do the implementation.
And here we are, modifying all to standard so we can ditch the " open a tiket for IPV6 /64 " and also this give us the following flexibility wise:
I know we are not winning loyalty points stopping your services, but we had 2 options on the table:
A. send out 500-600 E-mails so customers set up the new IPV6 setup or do a power off power on to auto apply the stings - all this when they read the email. - most probably this would taken 1 Decade
B. kill the old setup, we do all the work + some of you that have "nerd" setups - this will take ~1 week by our estimates.
I chose option B, and if you want to blame someone, blame me.
I really did not want to drag with us into 2025 an "improvisation" but rather start the year clean and focus on the more important stuff.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”
Sorry but I am still a bit confused.
From your previous offers:
As I did not request for a /64 previously, will my "1 Dedicated IPv6 IP" continue to work?
Or does it has to be replaced with a /64?