Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


racknerd will forcibly replace my IP (everyone in LA datacenter)
New on LowEndTalk? Please Register and read our Community Rules.

racknerd will forcibly replace my IP (everyone in LA datacenter)

linuxdevlinuxdev Member
edited September 20 in General

Just got the following email from racknerd. which upon reading reads like they will take away the public ip I use now and give me some new one.

I checked with support and they confirmed it's exactly what will happen.

Now I realize racknerd is cheap and I am personally ready to work around such issues and I am not too impacted... The only services I have there are personal junk,

But this seems that could be annoying to a lot of people that use racknerd for "anything production level" and that have to suddenly update DNS records (and maybe more - depending on how they use it), which probably means downtime as well.

thoughts?

Email:

We are reaching out to update you on an important IPv4-related change happening at our Los Angeles DC-02 location, which based on our records, one or more of your VPS services are currently hosted. If you have multiple services with us, please be aware that this only impacts KVM VPS services hosted out of our Los Angeles DC-02 location. All other datacenter locations of ours remain unaffected. To clarify, this change will only affect your IPv4 address; your data will remain intact, and entirely unaffected. IPv6 addresses also will not be changed.

Our upstream provider in Los Angeles DC-02 (Multacom) has initiated an IPv4 network renumbering project -- a change that is set to impact all downstream clients, including RackNerd. The renumbering process is pivotal for aligning with the long-term vision and expansion plans of the Multacom datacenter, which is now part of the EdgeCentres family.

Please note, that you will not experience any price changes, and additionally, we understand that our Los Angeles DC-02 location continues to experience remarkable popularity, notably among users in the US west coast, Oceania, and Asia. The top-tier network performance at our Los Angeles DC-02 location will continue as you have come to expect.

To give you a clearer picture, we operate over 500 hypervisors (also known as host nodes) at this specific location. To ensure a smooth transition, we'll roll out this update in a structured and phased manner. Here's a breakdown of the scheduled IP migrations:

The IP renumbering will be conducted in the following batches:

Node names beginning with LAXSSD1* to LAXSSD3* -- Oct 1 to Oct 15, 2023
Node names beginning with LAXSSD4* to LAXSSD6* -- Oct 15 to Oct 31, 2023
Node names beginning with LAXSSD7* to LAXSSD9* — Nov 1 to Nov 15, 2023
Node names beginning with LAX0* — Nov 1 to Nov 15, 2023

How do you know what node your VPS is on? Log into the SolusVM control panel located at https://nerdvm.racknerd.com/ and click on "Manage" on the VPS in question. Once you do, you'll see a table column for "Node". Here’s a video tutorial on how to do so.

The renumbering process will be automated but closely monitored by our Engineering team. You can expect to receive an email with the subject line "[RackNerd] New IP Address Assigned" once the process is completed for your VPS. This email will include your newly assigned IP address details. Alternatively, you can sporadically check SolusVM during the respective timeframe; if you notice an IP change, you'll find the new information there. Given the sheer volume of the VMs involved, this is a sizable operation and will be systematically carried out over multiple weeks. After the new IP is assigned, we will automatically utilize the "Reconfigure Network" feature by SolusVM to automatically update your server's operating system network configuration over to the new IP address. All said and done, there should be less than a few minutes of downtime per VM.

Just to be clear, this is only an IP address change - and your data/settings/etc will remain fully intact.

We sincerely apologize for any inconvenience, and we are quite confident that the long-term benefits will make it worthwhile. Once the Multacom datacenter is fully integrated with the EdgeCentres network, we're excited to tap into their extensive capabilities and various points of presence. For context, you may have seen that our LA DC-02 location has frequently been out of stock over the past year. This has been largely due to constraints at the Multacom datacenter, especially when it comes to available space and limited facility power, which have been limiting our ability to grow. Now that Multacom is becoming a part of EdgeCentres, we're optimistic these roadblocks will be cleared. This paves the way not just for our growth at LA DC-02, but also for expanding into new datacenter locations with EdgeCentres, which in turn means more locations for our customers to deploy services out of.

Your continued trust in RackNerd is highly appreciated, and our team stands ready to assist you with any questions you may have through this transition. If you have any special requests or require unique accommodations related to this change, please respond to this email within 7 days (the sooner, the better). We will do our utmost to understand your requirements and strive to accommodate them to the best of our ability.

Thank you for your understanding, cooperation, and continued partnership with us. Should you have any questions or concerns regarding this, please e-mail our Engineering team directly at: [email protected]

Thank You,

RackNerd Engineering
[email protected]
20+ Datacenters -> Dedicated Servers, Private Cloud, DRaaS, Colocation & VPS
https://www.racknerd.com/

«13

Comments

  • EthernetServersEthernetServers Member, Patron Provider
    edited September 20

    This is unfortunately something that can and does happen at any level (web hosting, VPS, dedicated, colo, etc) and isn't all that uncommon.

    The only real way to protect against it is to buy (note: not lease) your own IP space and announce that space through the provider of your choice (also known as BYOIP - Bring Your Own IPs).

    Kudos to Racknerd for the level of detail they've included here.

  • AllHost_RepAllHost_Rep Member, Patron Provider
    edited September 20

    With regards to the above post ^

    It's less of an issue if you buy services directly from a provider who runs their own network and was allocated the IP address space directly from the RIR. Whilst it'll certainly never eliminate the risk of an IP address change in future, buying directly from such provider will drastically reduce the chance of it being required.

  • avelineaveline Member, Host Rep
    edited September 20

    I think it's happening to all customers (e.g. Cloudcone) using Multacom's IP space.

    They sold almost all of their IPv4 addresses to Amazon a month ago and I'm wondering why the sold IP space were still announced under their AS.

    See below for more details:

    https://account.arin.net/public/transfer-log#NRPM-8.4IPv4
    https://whois.arin.net/rest/org/MULTA/nets

  • bdlbdl Member

    thoughts?

    DNS is awesome

  • I hate IP changes. Triples our headache.

  • coventcovent Member
    edited September 20

    Happens once in awhile, for other providers as well. I've learned through past that it's just best to take into account when configuring the server initially so that the new public IP can be reconfigured easily for any services running on the machine. DNS side is usually pretty easily done with scripting through API (if there's a lot of domains/subdomains that it affects)

  • @aveline said:
    They sold almost all of their IPv4 addresses to Amazon a month ago and I'm wondering why the sold IP space were still announced under their AS.

    Ah interesting. thanks for sharing.
    So whatever ips they (Multacom) still have left - is what's going to be used by racknerd?

    this is my sign from the universe to enable ipv6.

  • NeoonNeoon Community Contributor, Veteran

    Time to setup dynDNS if you are a racknerd or virmach customer.
    On the next IP change all you need to do, is click the reconfigure botton and you set.

  • dustincdustinc Member, Patron Provider, Top Host

    Our Los Angeles DC-02 facility continues to see sustained high demand, particularly from users across the US west coast, Oceania, and Asia. This location operates out of the Multacom datacenter (our upstream provider) located at 707 Wilshire Blvd., which is now part of the EdgeCentres family.

    Since Multacom joining the EdgeCentres family of brands, my team and I have had several in-depth discussions with EdgeCentres management, both virtually and in person, in Downtown LA. We're quite optimistic about their long-term strategy and their aggressive expansion initiatives. We believe that in the long run, these developments will translate into benefits for you, our end users, and will also pave the way for us to tap into new markets in the foreseeable future.

    Our clients and prospective clients may have noticed that our Los Angeles DC-02 inventory has frequently been sold out throughout the past year. The reason behind this has been Multacom's limitations in terms of facility space and power capabilities to support our rapid growth. Now, with Multacom becoming a part of EdgeCentres, these bottlenecks are expected to be addressed, allowing us to not only continue our expansion but also to maintain more consistent inventory availability in this popular location.

    An element of our upstream provider (Multacom’s) integration into EdgeCentres is the renumbering of IPv4 addresses and the consolidation of smaller IPv4 CIDR blocks that were holding up entire /24's. Given the network's scale, the IPv4 renumbering project will be implemented in multiple phases. While this does cause a minor inconvenience in the short term for all downstream clients of Multacom (including us and thus our end-users), we are confident that the long term benefits will outweigh the short-term challenges.

    I want to be completely transparent about the intricacy of this renumbering process. We're talking about managing more than 500 physical host nodes in just this location, not to mention the countless individual virtual machines that reside here. Our team is zeroed in on trying to make this as seamless as possible for our customers.

    To help ease the transition, we’re incorporating the following measures:

    Automated Reconfiguration: Once your new IP(s) have been assigned, we will automatically attempt to "reconfigure network" via SolusVM for your VPS, meaning most of you won't have to manually adjust any network settings within your OS level, unless you're on a custom OS.

    Personalized Email Notifications: Instead of simply telling our customers to refer to SolusVM to know their new IP, we'll send out individual emails with the new IP address for your VPS. This adds an extra layer of convenience for you.

    Clear Documentation: The email notification will also include the old IP address for easy reference, making the transition transparent.

    Consistent IP Count for Multi-IP Customers: If you have multiple IPv4 addresses assigned to your VPS, there’s no need to worry about any count-discrepancies during the transition. We'll assess the number of IPs currently allocated to your VM and ensure that you receive an equivalent number of new replacement IPs. This way, you won’t have to go through the hassle of requesting additional IPs post-renumbering.

    Billing System Updates: We'll update your service records in our billing/support system (WHMCS) with the new IP(s) to eliminate any confusion when managing your services or seeking support.

    Minimal Downtime: All said and done, we are talking about under 2 minutes of downtime on average per VM.

    This change will not result in any pricing changes to our clients. RackNerd remains steadfast in our mission to provide unparalleled reliability and top-notch service. Additionally, for those curious if there will be any network carrier related changes - the good news is that you can expect the same robust network performance (Multacom’s BGP blend) in this location to remain the same.

    Once we cross this bridge, we look forward to new horizons (additional datacenter locations and more consistent inventory availability in Los Angeles DC-02).

    We greatly value your ongoing support and the trust you've placed in RackNerd as your provider. Feel free to reach out if anyone has any questions!

  • @bdl said:

    thoughts?

    DNS is awesome

    Right up until the point it stops working :)

    For me, it usually it picks a point to mess up right in the middle of doing completely unrelated configuration changes. That's when you try and figure out where you screwed up in your configuration file.... you find 1-2 posts on stack overflow thinking "YES! I'm going to fix it!" just to click into the posts and see both of the questions are exactly the situation you're in, but they're unanswered questions...

    I've lived that nightmare more times than I care to admit. Whenever that happens now, the FIRST thing I check is DNS lol

    Thanked by 2bdl sebkehl
  • At least they inform prior and has a friendly and responsive support so you can get all the help you need if necessary. Far better than certain other provider where it happens like every quarter, no support, and when asked about it replies like how dare you complain about a cheap service.

  • jtkjtk Member
    edited September 20

    @linuxdev said:
    thoughts?

    An unfortunate, but not uncommon occurrence with many low-end providers. This is most common and most likely when your hosting provider leases addresses from an upstream or broker. Ideally this would be relatively rare. To minimize the occurrence of such renumbering events, select providers that have their own address space and have good, long track record at network stability.

    Note, often times you'll find these renumbering events often come with physical location changes. This is a little less common, but with VirMach for example, both cases have been very frequent and have greatly devalued their service for my purposes.

    However, depending on what one uses the host for, this may be a non-issue. It is most annoying when you have to change anything associated with a public address, such as DNS, certs, various monitoring, etc.

    In short, this is perfectly natural and more or less annoying depending on your use case and tolerance for such events. There are lots of providers that will practically never renumber you for as long as you have the service.

    Thanked by 2yoursunny MannDude
  • bdlbdl Member
    edited September 20

    @Don_Keedic said:

    @bdl said:

    thoughts?

    DNS is awesome

    Right up until the point it stops working :)

    For me, it usually it picks a point to mess up right in the middle of doing completely unrelated configuration changes. That's when you try and figure out where you screwed up in your configuration file.... you find 1-2 posts on stack overflow thinking "YES! I'm going to fix it!" just to click into the posts and see both of the questions are exactly the situation you're in, but they're unanswered questions...

    I've lived that nightmare more times than I care to admit. Whenever that happens now, the FIRST thing I check is DNS lol

    @dustinc said:

    To help ease the transition, we’re incorporating the following measures:

    Automated Reconfiguration: Once your new IP(s) have been assigned, we will automatically attempt to "reconfigure network" via SolusVM for your VPS, meaning most of you won't have to manually adjust any network settings within your OS level, unless you're on a custom OS.

    oof - you're sure that's a good thing? I've had Solus completely bork my custom /etc/network/interfaces in the past when support thought the network reconfig was a Good Idea [tm]. Much hilarity ensued.

    Thanked by 2Don_Keedic sebkehl
  • @covent said:
    I've learned through past that it's just best to take into account when configuring the server initially so that the new public IP can be reconfigured easily for any services running on the machine. DNS side is usually pretty easily done with scripting through API (if there's a lot of domains/subdomains that it affects)

    how would you script something like this where its semi-reliable? You are not talking about dyn dns here right?
    as dyn dns typically would mean certain downtime.

  • @Don_Keedic said:
    Right up until the point it stops working :)
    the FIRST thing I check is DNS lol

    That. Two stickers on my laptop: One says "It's always DNS" and the other "stackoverflow".

  • @aveline said: They sold almost all of their IPv4 addresses to Amazon a month ago

    seems right. That's why my shared IP started showing Amazon as Parent IP holder instead of Multacom

  • @bdl said:

    oof - you're sure that's a good thing? I've had Solus completely bork my custom /etc/network/interfaces in the past when support thought the network reconfig was a Good Idea [tm]. Much hilarity ensued.

    That's another fun one! I don't think I've ever had that "reconfigure networking" deal ever work out in my favor. Whatever is broke just ends up more broken.

  • coventcovent Member
    edited September 20

    @linuxdev said:

    @covent said:
    I've learned through past that it's just best to take into account when configuring the server initially so that the new public IP can be reconfigured easily for any services running on the machine. DNS side is usually pretty easily done with scripting through API (if there's a lot of domains/subdomains that it affects)


    how would you script something like this where its semi-reliable? You are not talking about dyn dns here right?
    as dyn dns typically would mean certain downtime.

    Check for example: https://developers.cloudflare.com/api/operations/dns-records-for-a-zone-update-dns-record

    I have a table of zones, ip addresses and domains which is then easy to iterate through with a script (if and when needed). So if a specific IP changes, I only modify it once in the table

    It's not only for cases when IP address changes in provider end but also if/when I decide to move a specific service to different host/provider. Makes managing multiple servers and domains much more feasible than trying to do everything manually in DNS dashboard

    Alternative approach, if you don't want to maintain a table:
    At least for Cloudflare you can export DNS records BIND config for whole zone, then find and replace IP addresses as needed and import modified config back in. All possible through API with a simple script (created with ChatGPT ;) )

  • @dustinc just IPV4 or also IPV6 ?

    Thanked by 1yoursunny
  • @Carlin0 said:
    @dustinc just IPV4 or also IPV6 ?

    From the quoted email;

    IPv6 addresses also will not be changed.

    Thanked by 1Carlin0
  • dustincdustinc Member, Patron Provider, Top Host
    edited September 20

    @bdl said:

    @Don_Keedic said:

    @bdl said:

    thoughts?

    DNS is awesome

    Right up until the point it stops working :)

    For me, it usually it picks a point to mess up right in the middle of doing completely unrelated configuration changes. That's when you try and figure out where you screwed up in your configuration file.... you find 1-2 posts on stack overflow thinking "YES! I'm going to fix it!" just to click into the posts and see both of the questions are exactly the situation you're in, but they're unanswered questions...

    I've lived that nightmare more times than I care to admit. Whenever that happens now, the FIRST thing I check is DNS lol

    @dustinc said:

    To help ease the transition, we’re incorporating the following measures:

    Automated Reconfiguration: Once your new IP(s) have been assigned, we will automatically attempt to "reconfigure network" via SolusVM for your VPS, meaning most of you won't have to manually adjust any network settings within your OS level, unless you're on a custom OS.

    oof - you're sure that's a good thing? I've had Solus completely bork my custom /etc/network/interfaces in the past when support thought the network reconfig was a Good Idea [tm]. Much hilarity ensued.

    Hi @bdl -- Thanks for this feedback. Over the years, SolusVM has made some improvements to the "Reconfigure Network" feature, and while we've found that this will work for the majority of our customers, we understand that it may not be perfect, specifically for customers who have custom OS/custom network configs. For those with any special requests or require unique accommodations, we are happy to help in the best way possible and ask that you email our Engineering team directly at [email protected] (before the specified timeframe for the respective batch - the sooner, the better) so we can do our best to understand your situation and provide a tailored accommodation.

  • dustincdustinc Member, Patron Provider, Top Host

    @Carlin0 said:
    @dustinc just IPV4 or also IPV6 ?

    Hi @Carlin0 -- IPv6 will remain unchanged :) If you do not already have an IPv6 address allocated on your LA DC-02 VPS, feel free to open a ticket to request it (if desired). There is no extra charge for that.

    Thanked by 1Carlin0
  • Did not receive any mail so far regarding this, but thanks for the update.
    I had wondered why there was some flapping yesterday, makes sense now!

  • dustincdustinc Member, Patron Provider, Top Host

    @sh97 said:
    Did not receive any mail so far regarding this, but thanks for the update.
    I had wondered why there was some flapping yesterday, makes sense now!

    Hi @sh97 -- The email mentioned in the OP is still in the process of going out to all DC-02 customers :) We have a large number of clients in this location, and the way WHMCS' mass mail tool works is it sends it out in small chunks, then takes a break, then sends out another chunk again, and so on (rather than all at once). At the current pace, it likely won't finish sending to all DC-02 customers until another 1-2 more days.

    Thanked by 1sh97
  • @dustinc said:

    @Carlin0 said:
    @dustinc just IPV4 or also IPV6 ?

    Hi @Carlin0 -- IPv6 will remain unchanged :) If you do not already have an IPv6 address allocated on your LA DC-02 VPS, feel free to open a ticket to request it (if desired). There is no extra charge for that.

    Could you please ensure PTR is automatically taken care of as well.
    That will be amazing.

  • MikePTMikePT Moderator, Patron Provider, Veteran
    edited September 20

    @aveline said:
    I think it's happening to all customers (e.g. Cloudcone) using Multacom's IP space.

    They sold almost all of their IPv4 addresses to Amazon a month ago and I'm wondering why the sold IP space were still announced under their AS.

    See below for more details:

    https://account.arin.net/public/transfer-log#NRPM-8.4IPv4
    https://whois.arin.net/rest/org/MULTA/nets

    Damn so many huge subnets! And many, many millions for sure.

  • @dustinc said: We're talking about managing more than 500 physical host nodes in just this location,

    Hard to believe. Are those owned or you rent them?

  • dustincdustinc Member, Patron Provider, Top Host

    @srch07 said:

    @dustinc said:

    @Carlin0 said:
    @dustinc just IPV4 or also IPV6 ?

    Hi @Carlin0 -- IPv6 will remain unchanged :) If you do not already have an IPv6 address allocated on your LA DC-02 VPS, feel free to open a ticket to request it (if desired). There is no extra charge for that.

    Could you please ensure PTR is automatically taken care of as well.
    That will be amazing.

    Hi @srch07 -- rDNS/PTR will continue to be supported on the new IPv4 ranges :) As for getting the record set on the new IP, it's best to request this after you update your DNS A records for the respective value to point to the new IP accordingly.

    Thanked by 2Carlin0 srch07
  • dustincdustinc Member, Patron Provider, Top Host

    @neckbreaker said:

    @dustinc said: We're talking about managing more than 500 physical host nodes in just this location,

    Hard to believe. Are those owned or you rent them?

    Hi @neckbreaker -- It is a mix, we have colocated infrastructure and leverage their hardware too. A fair number of our nodes are indeed rented. We have a very amicable and established relationship with Multacom and EdgeCentres, including with the original founders who are now in executive roles within the company.

    To confirm, we do in fact have 500+ physical servers in this location. Because of this, we're working extra hard to ensure a smooth transition.

  • @dustinc said:

    @Carlin0 said:
    @dustinc just IPV4 or also IPV6 ?

    Hi @Carlin0 -- IPv6 will remain unchanged :) If you do not already have an IPv6 address allocated on your LA DC-02 VPS, feel free to open a ticket to request it (if desired). There is no extra charge for that.

    Thank you, I already have 5 IPV6 ;)

Sign In or Register to comment.