<somethingfar[m]>
How do big server get "big hosts" in IPFS? Is it this way:
<somethingfar[m]>
If the server gets many requests for a specific hash, then it will (on its own decision when to share too) it as well download the file and pin it. Is this right?
<somethingfar[m]>
How do big servers get "big hosts" in IPFS? Is it this way:
<pcardune>
kind of. My understanding is that it will cache the response temporarily, but not pin it permanently.
<somethingfar[m]>
If the server gets many requests for a specific hash, then it will (on its own decision when to share too) it as well download the file and pin it. Is this right?
<Kubuxu>
no
<Kubuxu>
not currently
<deltab>
pcardune: I believe the usual promisify approach works
<pcardune>
deltab: it does, it's just kind of annoying to have to promisify every single api I plan to use. I'm wondering whether there is a specific reason why promises aren't just the default.
<deltab>
iirc, the thinking was something like "we could, but most interfaces aren't using promises yet, and people can promisify if they want"
<pcardune>
that sounds like a self-fulfilling prophecy :(
<pcardune>
The only reason I can imagine for not supporting both callbacks and promises is to save bytes... assuming they expect people to use this primarily in a web browser.
<pcardune>
I would happily submit a PR that adds support for callbacks. Just want to know whether it will be rejected outright for some reason I'm not aware of.
<pcardune>
deltab: thanks for pointing me to those. I had a hunch there had already been epic amounts of discussion about this. Now I'm wondering whether the same discussion has occurred for js-ipld
bwn has quit [Ping timeout: 240 seconds]
ONI_Ghost has joined #ipfs
bwn has joined #ipfs
<Kubuxu>
pcardune: js-ipld should use the same schema as js-ipfs, for uniformity (AFAIK).
<pcardune>
I see some uses of the pull-defer library, but at the very least, the implementation of put() and get() just support callbacks
bwn has quit [Ping timeout: 240 seconds]
ccii has quit [Quit: Leaving.]
dimitarvp has quit [Quit: Bye]
jesse22 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
revlulz[m] has left #ipfs ["User left"]
ccii has joined #ipfs
anoemi has quit [Remote host closed the connection]
jesse22 has joined #ipfs
ninja_8_u has joined #ipfs
<ninja_8_u>
hey all, does windows 10 show the external IP when executing "ipfs id" at all?
<ninja_8_u>
(I'm finding this works on Linux machines, but Windows 10 fails to find its external address 100% of the time on the same subnet at my house)"
<ninja_8_u>
I've tried: Enabling UPNP on router, disabling SPI, port forwarding 4001, disabling IPv6 on the router... all with ipfs daemon restarts every go.
<ninja_8_u>
And disabling Windows Firewall.
<ninja_8_u>
it does seem to swarm listen and swarm announce on the LAN ip just fine. It can even communicate with other machines on that subnet just fine
<ninja_8_u>
But nothing seems to get it the external ip in "ipfs id", unlike linux machines on the exact same subnet.
cwahlers_ has quit [Read error: Connection reset by peer]
cwahlers has joined #ipfs
<ninja_8_u>
Oh, and here's the kicker. "ipfs swarm peers" shows a whole bunch of working connections
pcardune has quit [Ping timeout: 276 seconds]
und0ck3d has quit [Quit: und0ck3d]
tarrence has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Mateon3 has joined #ipfs
Mateon1 has quit [Ping timeout: 256 seconds]
Mateon3 is now known as Mateon1
<Kubuxu>
ninja_8_u: interesting, I would expect it to work
<Kubuxu>
can you try running "IPFS_REUSEPORT=false ipfs daemon" linux, if it disappeared then it means that your external IP is because of reuseport which IIRC is not supported on Windows.
Neomex has quit [Ping timeout: 264 seconds]
ccii has quit [Remote host closed the connection]
<ninja_8_u>
Kubuxu: you might be right here. Giving it a couple minutes to confirm
<Kubuxu>
but from another side it should work with UPNP too, not sure why it doesn't
Steverman has quit [Ping timeout: 264 seconds]
tarrence has joined #ipfs
saki has quit [Quit: saki]
jaboja has quit [Remote host closed the connection]
ericxtang has quit [Remote host closed the connection]
ericxtang has joined #ipfs
<ninja_8_u>
Kubuxu: Disabling all firewalls to rule out more
<ninja_8_u>
I found something very interesting though:
<ninja_8_u>
Linux: When running with IPFS_REUSEPORT="false", ipfs 0.4.13 doesn't create entries in UPNP table on router (I made sure to clear that table and stop all competing processes in subnet before starting ipfs daemon)
<ninja_8_u>
Windows: When running "ipfs daemon" normally, UPNP entries show up in table on router. (Again, made sure to clear table and stop competing processes on subnet before starting test).
<ninja_8_u>
Neither case results in the external IP showing up in "ipfs id".
<ninja_8_u>
^^ I left forwarding off for this test too.
<ninja_8_u>
(i.e. no explicit binds to 4001, or anything really)
<Kubuxu>
Hmm, the lack of UPNP entries in case of IPFS_REUSEPORT=false are interesting.
ericxtang has quit [Ping timeout: 248 seconds]
<Kubuxu>
I wish I could give you a instructions for checking logging for UPNP subsystem but it is 4AM here and it might be a mess.
<ninja_8_u>
Policy on firewall on that Linux box was accept/accept for both ipv4 and ipv6
<ninja_8_u>
Kubuxu: If there's a link anywhere online, I'd happily read up on it!
<ninja_8_u>
So I guess this means that without NAT, and without REUSEADDR, it's impossible to determine the external IP.
<ninja_8_u>
Sorry, meant to say "without UPnP, and without REUSEADDR"
<Kubuxu>
yeah, roughly. Likely it would be possible using the mapping port we got and observable address (without port) but nobody has implemented something like this.
<ninja_8_u>
My router shouldn't even be exposing that subnet!
<ninja_8_u>
it's in an entirely different vlan, and inter vlan routes are 100% turned off :P
<Kubuxu>
hahah
<ninja_8_u>
I'm in 192.168.1.0/24, not 192.168.100/24 :)
<ninja_8_u>
there is a 192.168.100.1 on the other side tho:
<ninja_8_u>
Me <-192.168.1.0/24-> RV320 <-198.x.x.x ... because bridge mode-> ISP Router (Happens to have a 192.168.100.1 when you jump on its subnet).
<ninja_8_u>
but I can't access it at all from this side. I don't know how anything would even be able to route to it.
<ninja_8_u>
Sorry I lied, I'm not in bridge mode. The ISP router is in bridge mode.
<ninja_8_u>
RV320 is just doing a normal route, so, technically you could conceivably route to 192.168.100.1
<ninja_8_u>
let me nail this down, I want to get it right
Ekho has quit [Remote host closed the connection]
ninja_8_u has quit [Ping timeout: 256 seconds]
Ekho has joined #ipfs
ninja_8_u has joined #ipfs
}ls{ has joined #ipfs
K0HAX has quit [Ping timeout: 256 seconds]
colatkinson has joined #ipfs
K0HAX has joined #ipfs
<ninja_8_u>
lol
<ninja_8_u>
Okay, back... and yes, it is DEFINITELY leaking information from the different vlans
<ninja_8_u>
this is my own dumb fault for trying something smart xD
}ls{ has quit [Ping timeout: 260 seconds]
<ninja_8_u>
Kubuxu: ^^ :)
}ls{ has joined #ipfs
colatkinson has quit [Ping timeout: 268 seconds]
colatkinson has joined #ipfs
colatkinson has quit [Read error: Connection reset by peer]
BZ55 has quit [Ping timeout: 248 seconds]
tarrence has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
und0ck3d has joined #ipfs
dgrisham has joined #ipfs
guideline has joined #ipfs
robattila256 has quit [Quit: WeeChat 2.0.1]
robattila256 has joined #ipfs
und0ck3d has quit [Quit: und0ck3d]
colatkinson has joined #ipfs
colatkinson has quit [Read error: Connection reset by peer]