Jump to content

vAMSYS Downtime - 20 May 2023


Lukas Jankauskas

Recommended Posts

  • vAMSYS Platform Administrator

When?

Saturday, 20 May, 01:00    UTC-7
Saturday, 20 May, 02:00    UTC-6
Saturday, 20 May, 04:00    UTC-4
Saturday, 20 May, 09:00    UTC+1
Saturday, 20 May, 10:00    UTC+2
Saturday, 20 May, 11:00    UTC+3
Saturday, 20 May, 16:00    UTC+8
Saturday, 20 May, 18:00    UTC+10

What services will be affected?

All vAMSYS services and affiliated websites - this includes:

  • vamsys.io, VDS and Orwell
  • Community Forums
  • Vision by vAMSYS
  • All API responses, including statistics and json data
  • MyNextAirline, and other VA websites we host

How long will vAMSYS be offline for?

Ideally - 12 hours or less from system shutdown to all systems operational again. The offline period might be longer, might be shorter - to that end, we are exploring couple of creative options in hopes of minimizing the downtime and will let you know if our time estimate changes prior to the day - but once we start we cannot stop and will do our best to return vAMSYS back as quickly as we can - meaning, on the day in question - vAMSYS will be back as soon as we are done without any promises of ETA.

For some users the wait might be longer if we need to change DNS records - but we will do our best to avoid it and will try not to leave you at the mercy of your ISP or DNS provider.

Can I fly during the outage?

Due to the nature of work and rather unpredictable timeline we strongly urge all pilots to finish their flights prior to the start of the maintenance window and do not start new ones. There is a theoretical chance that you will be able to file a PIREP after the systems are back online, but they will be missing position reports and were your sim or Pegasus were to crash - you will not be able to resume tracking.

What are we doing and Why?

Our servers have been running for over a year now (1 year and 80 days at the time of this writing) and whilst they were receiving constant care and attention, the time has come to rebuild and retool our machines to enable us to effectively manage our infrastructure.

Our current setup does not allow us to easily redistribute resources (CPU and RAM allocations) between virtual machines - we used our best guess when we set them up a year ago and some of the numbers are off. The process of realigning our resources involves destructive action and hence the downtime - the web servers and databases cannot run when we move them about.

We've already started preparing - we have a breakdown of new Virtual Machine allocations, this time based on hard numbers gathered over the year. In practice, what this all means is that starting next week we'll shift the gear and start preparing for server wipe in earnest - this involves going through our VMs and noting down configurations, taking backups and all the other steps to get ready.

In the morning in question we'll take last database backups, zip up forum uploads and attachments and wipe the servers. Once done - we will start rebuilding 20 or so Virtual Machines which power vAMSYS ecosystem.

For those of you interested in more nitty gritty technical - we are keeping our two AMD Epyc 7642 servers. We are committed to lease them through to the end of Q1 2024 and we generally like them - they are powerful machines. We are tweaking one of the servers and upping it's RAM from 256 to 512 GB though - this change involves downtime, so now is the perfect time to do it. 

In addition, we will no longer use VMware virtualization tools, opting for Proxmox VE instead. It's just the right level of 'enterprise' to manage virtualization for us without the cost and complexity of it's predecessor. 

When all is said and done, we will be in a good position once again with infrastructure setup which is easier to manage and maintain which makes us happy campers.

  • Like 3
  • Thanks 1
Link to comment
Share on other sites

On 11/05/2023 at 11:04, Lukas Jankauskas said:

When?

Saturday, 20 May, 01:00    UTC-7
Saturday, 20 May, 02:00    UTC-6
Saturday, 20 May, 04:00    UTC-4
Saturday, 20 May, 09:00    UTC+1
Saturday, 20 May, 10:00    UTC+2
Saturday, 20 May, 11:00    UTC+3
Saturday, 20 May, 16:00    UTC+8
Saturday, 20 May, 18:00    UTC+10

What services will be affected?

All vAMSYS services and affiliated websites - this includes:

  • vamsys.io, VDS and Orwell
  • Community Forums
  • Vision by vAMSYS
  • All API responses, including statistics and json data
  • MyNextAirline, and other VA websites we host

How long will vAMSYS be offline for?

Ideally - 12 hours or less from system shutdown to all systems operational again. The offline period might be longer, might be shorter - to that end, we are exploring couple of creative options in hopes of minimizing the downtime and will let you know if our time estimate changes prior to the day - but once we start we cannot stop and will do our best to return vAMSYS back as quickly as we can - meaning, on the day in question - vAMSYS will be back as soon as we are done without any promises of ETA.

For some users the wait might be longer if we need to change DNS records - but we will do our best to avoid it and will try not to leave you at the mercy of your ISP or DNS provider.

Can I fly during the outage?

Due to the nature of work and rather unpredictable timeline we strongly urge all pilots to finish their flights prior to the start of the maintenance window and do not start new ones. There is a theoretical chance that you will be able to file a PIREP after the systems are back online, but they will be missing position reports and were your sim or Pegasus were to crash - you will not be able to resume tracking.

What are we doing and Why?

Our servers have been running for over a year now (1 year and 80 days at the time of this writing) and whilst they were receiving constant care and attention, the time has come to rebuild and retool our machines to enable us to effectively manage our infrastructure.

Our current setup does not allow us to easily redistribute resources (CPU and RAM allocations) between virtual machines - we used our best guess when we set them up a year ago and some of the numbers are off. The process of realigning our resources involves destructive action and hence the downtime - the web servers and databases cannot run when we move them about.

We've already started preparing - we have a breakdown of new Virtual Machine allocations, this time based on hard numbers gathered over the year. In practice, what this all means is that starting next week we'll shift the gear and start preparing for server wipe in earnest - this involves going through our VMs and noting down configurations, taking backups and all the other steps to get ready.

In the morning in question we'll take last database backups, zip up forum uploads and attachments and wipe the servers. Once done - we will start rebuilding 20 or so Virtual Machines which power vAMSYS ecosystem.

For those of you interested in more nitty gritty technical - we are keeping our two AMD Epyc 7642 servers. We are committed to lease them through to the end of Q1 2024 and we generally like them - they are powerful machines. We are tweaking one of the servers and upping it's RAM from 256 to 512 GB though - this change involves downtime, so now is the perfect time to do it. 

In addition, we will no longer use VMware virtualization tools, opting for Proxmox VE instead. It's just the right level of 'enterprise' to manage virtualization for us without the cost and complexity of it's predecessor. 

When all is said and done, we will be in a good position once again with infrastructure setup which is easier to manage and maintain which makes us happy campers.

Hi, cannot run the pegasus app anymor. just after the site maintenance.

Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...