MaGog is back to "normal" Legacy Mode.

Nice! Gonna report today's and yesterday's updates in a minute. :)

That's (actual) good news! At least as far as the limitations of that mode allow it to be...

* Added support for one new genre: JRPG

This morning my VPS service provider has upgraded its network, which should have translated to much faster performance. However, for reasons unknown, MaGog's update script (which harvests thousands of GOG pages and is therefore quite network-dependent) now runs about 3-4 times SLOWER (10-12 hours to complete, instead of 2.5-3 hours). This makes it impossible to run on a 6-hour cycle, as before. I have already aborted two cycles and have now disabled the 6-hour cron script.

I don't know the reason for this. I have contacted my VPS service provider but haven't heard from them yet.

Until I find a solution, I will run the MaGog update script manually once or twice a day, at irregular times. The next one should be finished in about 12 hours, fingers crossed.

I'll keep you updated (pun intended).

Updates breaking things... You'd almost think they're running Windows...
If it can't be certain to finish in under 12h to be able to set it to run every 12h, couldn't you make it run every, say, 16h, to not have to do it manually?

Yes, I can, but I want to closely monitor the next few runs to see if I can learn anything. If it persists, I'll set it to a longer automatic cycle.

Are they sure that they upgraded?

On a more serious note, have you noticed any differences (better or worse) with your other projects?

Don't know about MaGog's scripts, but the download speed for the database dump has certainly increased a lot. It used to download at 100 KB/s tops, and now it almost used up all of my DSL bandwidth. It's something... :)

Like muntdefems said, most everything is significantly faster, except for MaGog's update script.

Why doesn't your VPS service provider like our lady, why?!

I've seen this kind of thing before and I'd be careful whenever anyone mentions "network upgrades".

Yes, they will upgrade their equipment and landlines (switching to fiber optics and whatnot), and while the end result is a faster throughput for established connections, sometimes it also means you get a sluggishness in what I call burst transactions as part of the bundle.

I don't know who writes these new crazy hardware-level QoS optimizing routines, but they're assuming we're all in favor of prioritizing long-lasting high-throughput links and we're happy with our latency going to hell.

UPDATE:

No solution yet.

On Thursday, I got a not-too-unexpected response from my VPS service provider that read:
"Thousands of customers are updating their server right now, because of the IP changes, so network speed could be slower today.
Tomorrow you should see much faster times than previously. If you continue to see issues tomorrow let us know."

Forty-eight hours have passed and I see no improvement. The MaGog update script takes about 10 hours to run (instead of the usual 2.5 hours before).

I've contacted them again with some suggestions. Let's hope they look into it and not respond with another boilerplate ("use galaxy") reply.

Meanwhile, I've changed the cron scheduler to run the script only once a day. The script will begin to run at 12:30 Toronto time (18:30 Warsaw time) and the website should be updated about 10 hours later, around 22:30 Toronto time (04:30 Warsaw time).

If my VPS service provider manages to fix it soon, great. If not, I'll begin looking into moving my websites to another provider, but that would be a huge investment in terms of time and effort, so it may take me a while.

Sorry mrkgnao, I just want to let you know I accidentally sent you a friend request (was accepting one in Galaxy and clicked the wrong button). Please feel free to not accept it (unless you of course want to too). Apparently there is no way to cancel a pending request.

Again, my apologies.