Stability issues on multiple servers (new accounts most affected)

System Issues
1

Stability issues on multiple servers (new accounts most affected)

Since around the 2nd of November, we have seen stability problems on the website servers with Website IPs 185.27.134.33, 185.27.134.34, 185.27.134.55 and 185.27.134.57.

This issue was caused by an error in the account provisioning system, resulting in too many new accounts being created on these servers. This resulted in the servers being overloaded.

New, faster hardware has been setup that should be more capable of handling this load, and data is being migrated to these new servers. Unfortunately, the old hardware is not powerful enough to both migrate the data and support regular operation. Because of this.

Which components are affected by this issue

The website servers with IP addresses 185.27.134.33, 185.27.134.34, 185.27.134.55 and 185.27.134.57 are affected by this.

You can find your Website IP in the client area in the Account Details.

What issues may you see

If your website is on one of these IP addresses, you may see the following errors:

  • ERR_CONNECTION_REFUSED
  • 502 Bad Gateway
  • Error 520: web server returns an unknown error

When will this issue be resolved

It may take between a few days and a few weeks for this issue to be resolved.

What can you do to fix this

If you don’t do anything, the server load will automatically return to normal in a week or two.

If you don’t want to wait for that (or want to help reduce the load), you can create a new hosting account, and migrate your website to the new account. This new hosting account will be created on a different server that doesn’t have these stability issues. And in doing so, you are helping reduce the load for others on the server.

We are aware of the issue and are doing our best to get this resolved as soon as possible.

I completely understand that this may cause confusion and you will have additional questions. I am happy to provide any information I have to help you.

However, to keep the contents of this topic clean, clear and concise, I ask that you carefully read the Outage Communication Rules before posting.

In a nutshell:

  • Please keep the posts about the outage in this topic, and all posts not about this outage outside of this topic.
  • Please post relevant questions only that have not been answered yet. It may be compelling to ask obvious questions like “when will this be fixed”, but basic information like that will be put in this message as soon as we know it.
  • Please carefully check the known symptoms and affected components.
    • If you are on the affected components, you are almost certainly affected by the outage. If so, please refrain from simply stating you are affected, it does not help get your service back up faster.
    • If you are not on the affected components but do experience all of the symptoms, please say so below so we can check if our list is incomplete.
    • If you do not experience all listed issues and are not on the components that are known to be affected, the issue you are experiencing is most likely not related to this outage. If so, please create a new topic in the Hosting Support category and describe the issue there.
  • If you say you are affected by the outage, please always include your website URL. I can’t check if you are affected by the outage if you don’t say which of your websites is affected.

Posts that do not follow these rules may be deleted without warning. This is to help ensure that all the relevant information is easily available to everyone reading this topic.

4 Likes
6

If I understood correctly, (and please correct me if I am wrong), you will let us bleed for 2 weeks, expecting that people will leave the server on their own will, so things can go back to normal.

I’d be interested to know more details on the part it says: “doing our best to get this resolved as soon as possible.”, and a more precise estimation of when we can expect this to be fully solved.

I personally don’t think that relying and waiting on user churn is a reasonable course of action.

7

Do you have any better ideas?

To fix this sooner, we would have to migrate a large number of accounts to different servers. However, this migration itself would put a lot of stress on the infrastructure and make things worse for everyone right now. This would also result in the migrated accounts getting new IP addresses, which may cause some downtime due to DNS caching, or even require action from the website owner to bring the website back up if they are using their own nameservers.

Besides, since this issue primarily affects new accounts, many websites will be new and don’t have a lot of content worth migrating in the first place. If you just installed a script and don’t have any content in it, you could just get a new account, move your domain to it and reinstall the software.

Again, as I wrote in the post, you don’t have to wait for the issue to solve itself. You can move your site to a different server if you want to.

Also, for load to normalize we don’t necessarily need people to leave, just not allocating any new accounts will help too. Many accounts generate the most load in the first few days. These are the times when people set up their website, install scripts and plugins, import data, and so on. Once the site is set up, load is reduced a lot. So by simply not putting any new accounts on the server for a while, load will go down as well.

Like I wrote: as soon as possible. And because relying on churn is the most sensible course of action, this can take 1-2 weeks.

Because churn is not a process we have strict control over, I cannot provide a more precise estimation than this.

5 Likes
8

Thanks for the answer, the only thing that comes to my mind is to then lockout “new accounts”, delete them and push them to new servers.

However, this migration itself would put a lot of stress on the infrastructure and make things worse for everyone

I am not sure that a migration would make things worse. At least for me, my site has been offline since some time already. So it can’t get worse here.

9

No, it would not be worse for your site, but worse for others sites that are working properly. The less damage done, the better.

You can move your domain to a new account and all will be fine, but if we try a mass-migration, there will be other consequences

6 Likes
12

Hi, I am in one of the mentioned servers so I think I am affected by the outage.
On the other hand Google is constantly informing me that my robot.txt can’t be reached and I guess the outage can be the reason.
My url is https:/www.terapiapsicologicaonline.com
Thank kou.

In fact, I may not wait and migrate my site to a new hosting account as stated above.

I did so, migrated it to a new account. I hope Google can index it now.

13

Hello. I have a website in the affected IP range that has now been out of action for over a week, I understand that there has been stability issues, but the issues seem just as bad and have not got any better.

Could you please provide an update on when this is likely to return to a normal service? I also wonder if you have any “how to” information on transferring my wordpress site (together with ssl certificates) to a new account please?

Many thanks

1 Like
14

  1. Backup your website
    How to backup your website

  2. Remove your domain from your old account

  3. Create a new account

  4. Import the data
    How to migrate your website to InfinityFree

5 Likes
15

Thank you. I have migrated my site to a local installation successful;y but was just a little worried about de-registering the domain - I thought I may be able to “transfer” it somehow. Thanks, I shall see how I get on :slight_smile:

16

As long as the domain is using our nameservers, you can just remove it from your old account and add it
to your new account right away. There is no “transfer” button (and we’re not going to add it because I know everyone would assume that the website would be transferred too, which just isn’t feasible), but doing it by hand is almost as fast.

3 Likes
19

So the estimate for the issue to be resolved was “a week or two” as of 18 days ago, and I’m still getting errors. Can we get an update? Is it actually improving at all?

1 Like
20

Well, since no update has been posted, we can infer that this issue is still ongoing. As for “is it improving”, I personally don’t know nor am I sure if Admin does. You can monitor this topic for updates, or if you want to transfer to a new account, you can do that too.

6 Likes
21

This issue is getting worse not better. At this time it is essentially unusable as a hosting platform.
Would prefer not to have to move to another hosting platform but might have no choice if this isn’t resolved asap.

Can we please have an update on when this will be resolved.
These problems don’t just sort themselves out and asking your users to migrate to a new account to fix the issue is negligent.
What is being done to resolve this?

2 Likes
22

Please see:

We don’t know when this will be resolved, as it will only be fixed once enough people move their accounts over to stables servers.

I mean we can’t just add more storage and resources to a server that is currently in-use without making the problem worse, can we? And this is free hosting, so even if we could do that, who is going to pay for the better hardware?

Can you explain why? If we move users sites over manually, everyone who is using an external DNS provider will have a comply broken and unusable site. We are not going to do something that will create as many problems as it solves.

There is nothing that we can do to solve this, only the users on the affected servers can help solve it by moving there sites to unaffected servers.

Majority of websites can by transferred to a new server in just a few hours, and that is certainly faster then waiting around for others to do it first.

3 Likes
23

That is a shocking response.
If it’s part of your business model to offer free hosting, if that involves incurring additional costs to deliver a stable service then you should have the capacity to do so. You absolutely should be able to add more storage to any of your servers if required.
Again, negligent in offering this service and then not being able to manage and support it reliably.
This dismissive response only supports that.
It has not been made clear that the “only” solution was for users to move to a new account.
And expecting all your effected users to spend the “hours” doing so is again negligent.

Even if I wanted to migrate to a new account, currently FTP connections to the server are failing so I can’t even take a backup. So your only solution isn’t even an option right now.

Interesting that your pinned post underneath this one is asking people to submit a trustpilot review.
Happy to fulfil that request.

24

Then please take this as a lesson to always have a backup. Even if you are on paid hosting, hardware fails, and things happen. If you don’t have a backup, you really don’t own your own files, you are just using them.

And also note that adding and changing hardware on a production server is not an easy task. After buying the hardware and waiting for it to arrive, installing it often results in downtime, ranging anywhere from a few hours to a few weeks. So we really don’t want that either.

5 Likes