Brief(ish?) Outage Coming Soon
-
I need to do a snapshot, destroy, create in order to get the disk space increased.
This will take, oh who knows how long.... but, I'm running it on the old instance now as a test. I'm hoping to get started at 4PM Eastern.
-
Good luck
-
9PM BST (8PM UTC) for anyone wondering.
-
crossing my fingers!
-
We're back - does this mean it's done? @apapadimoulis
-
I thought that earlier; it then went down again, and when it came back, a few posts had been lost (only about three, and none of them worth anything).
I'm guessing that's just how the process went…
-
Yes, seems that way. Apparently after creating the snapshot, it turned the server back on...
Anyway, new droplet, with 60GB instead of 40GB now.
-
MOAR SPACE!!!
I promise not to fill it up with pictures of hedgehogs ;)
-
I saw server cooties times around 125 SECONDS
Thats pretty extreme... (i mean that i was answering at all)Also might wanna remove the banner
-
I saw server cooties times around 125 SECONDS
that's apparently how long nodejs is willing to wait for a connection.
;-)
-
It's also close to the max TCP/IP timeout of 120 seconds, IIRC
-
hmmm... i'm not all that familiar with network stack, but i though TCP datagrams specified TTL in hops, not absolute time?
-
TTL is basically to stop the packets getting caught in infinite routing loops, so they don't clog up the tubes
But then, my two-minute timeout assumption is balls anyway, as it appears to be a configurable setting on modern OSes
-
ahh, so you're telling me that you got no response, and just set response time to ~128 sec. That's not entierelklmdmkld informative...
http://i.imgur.com/bY5BQV5.png
-
Well, not quite; it's how long
node
waited for the request to return before saying 'ah, sod it!' and giving up ;)
-
So he's right - there was no response but it's down as responding in like 130 seconds.
-
He's right, but for the wrong reason; I gave him the right reason
-
ahh, so you're telling me that you got no response, and just set response time to ~128 sec. That's not entierelklmdmkld informative...
hmm.... yeah. that's why there's another graph forthcoming that uses the discoappdex scores. it's more informitive but also more boring.
in the mean time i refer you to this issue: https://github.com/SockDrawer/SockSite/issues/21
-
Did the outage happen?
-
Yes; we now have an extra 20GB for you to fill with Dwarf Fortress screenshots and snippets of Go
-
we now have an extra 20GB for you to fill with Dwarf Fortress screenshots and snippets of Go
Belgium it, hedgy!
-
Yes; we now have an extra 20GB for you to fill with Dwarf Fortress screenshots and snippets of Go
Just don't let him post in lojban anymore.
-
TTL is the maximum lifetime of an individual packet in hops. It's a field in the IP header.
The TCP connection as a whole has a number of separate timeouts in seconds, indicating how long a client must wait for packets to come in before dropping the connection, as well as how long a client must wait after a connectioon has been torn down before a new connection may be opened to the same destination on the same port pair.
-
hu-uh.....
TIL more about networking than i knew before.
-
How does this plot even work with the plot curves turning back on the X axis?!
-
It's the smoothing function that's trying to connect the dots with nice curves. It's not really a continuous graph.
-
How does this plot even work with the plot curves turning back on the X axis?!
that's a spline graph... does weird things with discontinuities. I'm switching to straight line for a future version.
-
-
I like the backwards lines. Keep it. It's so consistent.