chuan has quit [Remote host closed the connection]
<asheesh>
I am happy to announce that https://sandcats.io has a new cert that is valid for about a year!
<asheesh>
by which I mean, yay, I'm done dealing with that possibly-stressful task!
chuan has joined #sandstorm
<asheesh>
Howdy chuan!
<chuan>
hi!
<asheesh>
Let us know if know if we can help you with anything, or if you just want to chill quietly (aka idle).
* chuan
continues lurking :)
KooBaa has quit [Remote host closed the connection]
jadewang has quit [Remote host closed the connection]
jadewang has joined #sandstorm
NwS has quit [Ping timeout: 244 seconds]
ocdtrekkie has quit [Remote host closed the connection]
aldeka has quit [Remote host closed the connection]
KooBaa has joined #sandstorm
jadewang has quit [Remote host closed the connection]
frann has joined #sandstorm
OOCoder has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Remote host closed the connection]
KooBaa has quit [Ping timeout: 264 seconds]
bencevans has joined #sandstorm
sydney_untangle has quit [Read error: No route to host]
sydney_untangle has joined #sandstorm
kecors has quit [Quit: Leaving]
bencevans has quit [Quit: ZNC 1.6.1 - http://znc.in]
bencevans has joined #sandstorm
bencevans has quit [Changing host]
bencevans has joined #sandstorm
bencevans has quit [Remote host closed the connection]
bencevans has joined #sandstorm
bencevans has quit [Changing host]
bencevans has joined #sandstorm
Triplecute has quit [Ping timeout: 276 seconds]
Triplefox has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 240 seconds]
<dwrensha>
So, one gotcha with dropping in a new sandstorm binary for an existing install:
<dwrensha>
make sure to turn off auto-updates first!
<dwrensha>
otherwise, every ten minutes the autoupdater will run and restart Sandstorm
<dwrensha>
curiously, the autoupdater downloads Sandstorm and unpacks it, but then does not replace the existing install, because the version number on the existing directory is the latest version number
<dwrensha>
I suppose this is what I get for copying a binary between machines and dropping it in-place into an existing install, rather than using a supported method like `make update`.
<kentonv>
dwrensha: good thing our stats don't count the after-10-minutes update
<dwrensha>
one unfortunate side effect of me moving apps over to vagrant-spk is that it makes it more difficult for me to test new `spk` and `sandstorm-http-bridge` features
<dwrensha>
install-sandstorm-deps.sh or whatever could fix that
jadewang has joined #sandstorm
KooBaa has joined #sandstorm
jadewang has quit [Ping timeout: 260 seconds]
_whitelogger_ has joined #sandstorm
KooBaa has quit [Ping timeout: 244 seconds]
_whitelogger has quit [Remote host closed the connection]
jadewang has quit [Remote host closed the connection]
jadewang has joined #sandstorm
jadewang has quit [Remote host closed the connection]
KooBaa has joined #sandstorm
KooBaa has quit [Remote host closed the connection]
<dwrensha>
I still haven't figured out which process exactly is using the value from TMPDIR
<dwrensha>
The Ruby stuff uses the tmp directory relative to the root of the project, and fiddling with that did not fix the problem.
KooBaa has quit [Ping timeout: 240 seconds]
<asheesh>
"obviously" you should write an LD_PRELOAD that incorporates the PID into the value of TMPDIR, calculated at runtime by patching getenv("TMPDIR")
<dwrensha>
in any case, I would be +1 on having vagrant-spk's launcher.sh do the TMPDIR=/var/tmp thing as in my commit above
synchrone has quit [Ping timeout: 248 seconds]
<dwrensha>
asheesh: that sounds fun, actually. You are tempting me to learn more about LD_PRELOAD.
<maurer>
dwrensha: what do you want to know about LD_PRELOAD
<dwrensha>
I want to emulate tty devices inside Sandstorm grains.
<asheesh>
/usr/bin/yes
<asheesh>
I want that a lot personally also.
<asheesh>
dwrensha: I personally tihnk that it should be "pretty easy" to make that happen by copy-pasting the kernel's C code.
<asheesh>
(naturally subject to GPL attribution/etc requirements)
<dwrensha>
interesting... with my new spk binary running in a virtualbox vagrant-spk environment, 70% of the cpu time is at the symbol "_raw_spin_unlock_irqrestore"
<dwrensha>
above it on the stack is "__write_nocancel"
<asheesh>
I wonder if this gets fixed if you decrease the # of CPUs provided to Vagrant by 50%
<asheesh>
Specifically I wonder if this is hyperthreading-related horror.
<dwrensha>
looks like vagrant-spk sets the number of cpus to be equal to the host's number of cpus?
<asheesh>
Yeah, I am just vaguely fearful of hyperthreading.
<asheesh>
I can be more specific:
<asheesh>
I'm concerned that the Linux kernel within VirtualBox maybe doesn't know which "CPUs" are connected to each other in a hyperthreading way, and so schedules things poorly.
<asheesh>
Along those lines, one "easy" way to find out if this is the problem is to run VirtualBox with just one core, and see if things are faster.
<dwrensha>
I'll try with 3 cores
<asheesh>
If so, maybe it makes sense to expose just 1 CPU to the guest from vagrant-spk, even though the Sandstorm vagrantfile would use 1 core.
<dwrensha>
which is how many I give my usual virtualbox vm, which does seem to perform better than the vagrant-spk one
<dwrensha>
(fwiw, `sysctl -n hw.ncpu` returns "8" for me)
<asheesh>
Thoughts from elsewhere, this time w/r/t "turbo boost": "Basically this says that if you have processors sitting 50% or so idle in a guest, it may be prudent to trim back on your core count to the guest." <https://forums.virtualbox.org/viewtopic.php?f=7&t=31770>
<asheesh>
So yeah, I guess I would try with cores=1 (or cores=3, sure) at some point.
rustyrazorblade has quit [Quit: rustyrazorblade]
synchrone has joined #sandstorm
prettyvanilla_ has quit [Quit: Konversation terminated!]
prettyvanilla has joined #sandstorm
synchrone has quit [Ping timeout: 252 seconds]
notevil has quit [Read error: Connection reset by peer]
neynah has joined #sandstorm
xet7 has quit [Quit: Leaving]
Purity has joined #sandstorm
xet7 has joined #sandstorm
<Purity>
Hi!
<dwrensha>
Hi!
Purity has left #sandstorm [#sandstorm]
notevil has joined #sandstorm
KooBaa has joined #sandstorm
<maurer>
kentonv: Now that my isCanonical() function appears right, I was going to try my hand at canonicalize(). However, it seems like I'd either need to modify WireHelpers::copyPointer with a flag (to get it to do things like proper list and struct truncation when the flag is flipped), or create a separate set of routines doing this in PointerBuilder, StructBuilder, and ListBuilder. Which would you find more
<maurer>
tasteful?
mnutt__ has joined #sandstorm
KooBaa has quit [Ping timeout: 250 seconds]
notevil has quit [Quit: +++ ATZ]
asmyers has joined #sandstorm
amyers has quit [Remote host closed the connection]
asmyers has quit [Ping timeout: 244 seconds]
KooBaa has joined #sandstorm
Jan\ has quit [Ping timeout: 268 seconds]
wolcen has quit [Ping timeout: 250 seconds]
Jan\ has joined #sandstorm
KooBaa has quit [Ping timeout: 244 seconds]
dwrensha has quit [Ping timeout: 276 seconds]
wolcen has joined #sandstorm
dwrensha has joined #sandstorm
<kentonv>
maurer: Probably a flag to copyPointer() makes sense.