Emergency Site Maintenance

[RichM]

We're upgrading Canal World to a new and improved server next week. This upgrade is necessary because Canal World has been using more memory lately. The new server will have double the memory capacity, more storage, and more powerful CPUs to make sure Canal World runs smoothly and can handle its increasing demands.
 

We will try to keep any disruption to a minimum, though the site will be offline for for a few hours one evening during the migration. I don't have a specific time for the migration yet, but will keep you posted.
 

Apologies for any inconvenience. 

[MtB]

11 minutes ago, RichM said:

We're upgrading Canal World to a new and improved server next week. This upgrade is necessary because Canal World has been using more memory lately. The new server will have double the memory capacity, more storage, and more powerful CPUs to make sure Canal World runs smoothly and can handle its increasing demands.
 

We will try to keep any disruption to a minimum, though the site will be offline for for a few hours one evening during the migration. I don't have a specific time for the migration yet, but will keep you posted.
 

Apologies for any inconvenience. 

 

 

Is CWF traffic actually increasing to that extent then? 

 

Or is the upgrade necessary due to the recent Invision (bloatware?) upgrade?

[RichM]

1 hour ago, MtB said:

 

 

Is CWF traffic actually increasing to that extent then? 

 

Or is the upgrade necessary due to the recent Invision (bloatware?) upgrade?

The recent Invision update was just bug fixes and removal of some deprecated admin features.

 

We have to update other software too, not just Invision. This includes things like web server software, database software, anti virus software, mail server software. - Plus the operating system itself (Ubuntu) The list goes on... Over time, they consume more resources which is unfortunately the nature of software development but luckily hardware technology advances too. - though at a cost. Ultimately there becomes a point where we have to upgrade. Our current web server has lasted us over 3 years which seems reasonable. If we don't keep on top of this the site will be become slow, unreliable, glitchy and potentially insecure. We follow best practices to avoid this.

 

Also, site activity has increased in recent months, particularly the number of concurrent online users during peak times. I haven't drilled deep into the stats, but definitely an observation - this too will result in higher memory utilisation. 

This is at an increased cost to Canal World which as you know is funded entirely from member donations. All donations appreciated. 🙂

[RichM]

CWDF will be offline for a few hours this evening (after 21:00 GMT) for maintenance. Updates to follow.

[Jerra]

Thanks for the heads up and as always many thanks for your efforts on our behalf.

[RichM]

The migration to the server is now complete and took just under an hour. The new server doubles our memory capacity and gives us additional storage to accommodate things like attachments etc. It will also ensure we have enough resources to meet demands for the foreseeable future. 

 

The new server does unfortunately come at a greater cost to the site, which is funded entirely by our members. We're grateful for all donations received. Feel free to make a donation by clicking here if you wish.

 

Should you encounter any glitches, please first clear your browser caches. While it doesn't fix all problems, it will fix many. Instructions on how to do this can be found linked below:

https://helpcenter.trendmicro.com/en-us/article/tmka-20816

It is possible the site may take a little longer to come back online for some users due to a pesky problem known as "DNS Propagation". We have taken steps to mitigate this but can't avoid it altogether. - there are some tricks to help speed this up but the good news is that it will sort itself out in a short while. If you're reading this, chances are you don't have this problem.

 

Any other issues or questions, please let us know. 

[tree monkey]

14 hours ago, RichM said:

The migration to the server is now complete and took just under an hour. The new server doubles our memory capacity and gives us additional storage to accommodate things like attachments etc. It will also ensure we have enough resources to meet demands for the foreseeable future. 

 

The new server does unfortunately come at a greater cost to the site, which is funded entirely by our members. We're grateful for all donations received. Feel free to make a donation by clicking here if you wish.

 

Should you encounter any glitches, please first clear your browser caches. While it doesn't fix all problems, it will fix many. Instructions on how to do this can be found linked below:

https://helpcenter.trendmicro.com/en-us/article/tmka-20816

It is possible the site may take a little longer to come back online for some users due to a pesky problem known as "DNS Propagation". We have taken steps to mitigate this but can't avoid it altogether. - there are some tricks to help speed this up but the good news is that it will sort itself out in a short while. If you're reading this, chances are you don't have this problem.

 

Any other issues or questions, please let us know. 

As always thank you for your work on behalf of the site and its members 

[StephenA]

Glad you got it all done - it's never easy doing a server upgrade. Luckily when I was upgrading the Canalplan Server I was able to practice the build before we got the new server and once I've built it I kept syncing things up and actually kept both servers in sync for a few days before and after doing the final cut over.

[RichM]

On 19/02/2024 at 19:35, StephenA said:

Glad you got it all done - it's never easy doing a server upgrade. Luckily when I was upgrading the Canalplan Server I was able to practice the build before we got the new server and once I've built it I kept syncing things up and actually kept both servers in sync for a few days before and after doing the final cut over.

 

Keeping both in sync is definitely a clever way of doing it, maybe next time.

 

We have a fair amount of database activity so ideally I wanted to keep it simple and limited it to one sync from A to B. I reduced DNS TTLs right down to mitigate the impact of DNS propagation in between and had to block all user traffic at a firewall level to free up resources so I could do a SQL dump and compress the asset data. (we were running between 90%-95% memory utilization for a short period leading up to the migration!) Perhaps not the most elegant or seamless approach but it did the job without a single issue to report. (yet 🙂)