this post was submitted on 22 Jun 2024
158 points (98.8% liked)

Programmer Humor

32373 readers
555 users here now

Post funny things about programming here! (Or just rant about your favourite programming language.)

Rules:

founded 5 years ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] skilltheamps@feddit.de 1 points 4 months ago (2 children)

With something like this, how do you handle the period of time while copying? I mean you can't really leave it running as it wouldn't be in a consistent state. A "under maintenance" page instead? Copy to a fresh folder and when done tell the webserver to serve the new location?

[–] folekaule@lemmy.world 5 points 4 months ago* (last edited 4 months ago)

It depends. I've done it a few different ways:

  • YOLO: especially with thugs like PHP you only affect one page at a time and with low traffic the odds of a problem is small
  • Maintenance page: temporarily show a page. Some servers like IIS have this built in. Otherwise it's a simple update to httpd conf
  • In a cluster environment, just take the node you're updating out of rotation, and only update one node at a time.
  • Copy and switch like you suggested. Can be combined with any of the above and is a smart move if upload is slow or can be interrupted, or it's cumbersome to restore the old files

Edit: spelling

[–] jlh@lemmy.jlh.name 0 points 4 months ago (1 children)

Typically either downtime or taking the server off of the load balancer, depending on scale.

All of this is automated in modern container orchestration tools like kubernetes, though.

[–] skilltheamps@feddit.de 0 points 4 months ago

Thanks, yeah I'm well aware that this is not a problem in proper setups. I was just wondering what a company that is large enough to care about downtime and at the same time has staff manually click around in filezilla would do.