<isd>
Hey -- so my sandstorm instance's node process is taking up almost 100% of the system's memory.
<isd>
This has happened once or twice before, this time I swear I'll write it down: how do I take a dump of the heap for bug-reporting purposes?
<kentonv>
there's an undocumented meteor method... lemme see
<kentonv>
JS console: Meteor.call("heapdump")
<kentonv>
must be logged in as admin
<kentonv>
will show up in /opt/sandstorm/var/log/$TIMESTAMP.heapsnapshot
<kentonv>
it's taking 100% of system memory... but not 100% of CPU?
<kentonv>
(isd)
<isd>
Yeah, that sounds familiar.
<isd>
And actually I'm looking and it's more like 75%.
<isd>
But yeah, in the past there's been high CPU, but not this time.
<kentonv>
there's some sort of leak I've been trying to track down for a while
<isd>
...and there it goes. 100% cpu
<isd>
...and it's back to normal cpu now.
<kentonv>
yeah it's not actually in an infinite loop or anything, it's just that GC takes up more and more of CPU over time
<isd>
Ah.
<isd>
Anyway, I have a couple dumps now.
<kentonv>
until it gets to a point where there's not enough CPU to handle the incoming requests
<kentonv>
and then it falls over
<isd>
yeah, that's roughly what I've experienced.
<kentonv>
TBH I'm not entirely sure if I'm set up to use a dump from the Sandstorm (as opposed to Blackrock) build.
<isd>
Questions (a) is there anything I should do to sanitize this re: sensitive data? and (b) how can I make this of use?
strugee has quit [Ping timeout: 240 seconds]
<kentonv>
actually wait, heapdumps are totally self-contained IIRC
<kentonv>
they load into Chrome's dev tools
<kentonv>
so I guess you could load it up and see if anything jumps out
<kentonv>
I haven't actually gotten a chance to look at a heapdump for this issue yet
<kentonv>
so it's entirely possible
<isd>
It was fluctuating between 40 and 75%; became unusable at one point at which prompted me to look
<isd>
I'll upload it and send you a link
<isd>
I actually got like 3 of them.
jemc has joined #sandstorm
strugee has joined #sandstorm
<isd>
(send you a link via pm)
harish has joined #sandstorm
isd has quit [Quit: Leaving.]
pie_ has quit [Ping timeout: 240 seconds]
funwhilelost has joined #sandstorm
jemc has quit [Ping timeout: 245 seconds]
frigginglorious has quit [Ping timeout: 240 seconds]
jemc has joined #sandstorm
dwrensha has quit [Quit: Lost terminal]
jemc has quit [Ping timeout: 264 seconds]
dwrensha has joined #sandstorm
<ocdtrekkie>
So, apparently there's an awesome new security risk at clicking on ANY Google Ads, and it seems like nobody wants to know it exists.
<ocdtrekkie>
The top results for "youtube" and "amazon" have both been taken over today by Google ads that showed the official URLs in the destination, but went to malware sites in India when clicked.
<ocdtrekkie>
I found the behavior a full week ago though.
<TimMc>
lovely
<TimMc>
Probably no one is talking about it because most of us nerds have adblock.
<TimMc>
hmmm...
<TimMc>
ocdtrekkie: Are you *absolutely sure* you don't have malware on your machine? A bad browser extension, maybe?
<TimMc>
'cause that's a thing too, ad hijacking
<ocdtrekkie>
Absolutely. This has been tested in multiple browsers on multiple machines, personal and business, on different networks on different ISPs.
<ocdtrekkie>
And others on HN in small batches have been reporting and replicating the same behavior.
<ocdtrekkie>
Heck, I use Edge. Edge has almost no extensions to begin with, and they only pretty much can be installed from the Windows Store.
<ocdtrekkie>
But I've had Linux users test this as well, and of course, it's tested in Chrome, Firefox, Edge, and IE.
<ocdtrekkie>
When I contacted Best Buy about their ad being hijacked on Google Search, they confirmed they had reported the issue as well.
<TimMc>
Welp.
<TimMc>
Probably not malware then. :-P Worth asking, though.
<ocdtrekkie>
I know better than to click on them, mind you, my bigger concern is everyone I've specifically told, "Hey, check the URL line before you click to make sure it's the real site".
ill_logic has joined #sandstorm
jemc has joined #sandstorm
harish has quit [Ping timeout: 255 seconds]
funwhilelost has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
harish has joined #sandstorm
jemc has quit [Ping timeout: 240 seconds]
jemc has joined #sandstorm
gambatte has quit [Ping timeout: 240 seconds]
harish has quit [Ping timeout: 276 seconds]
FredFredFred_ has joined #sandstorm
FredFredFred has quit [Ping timeout: 240 seconds]
harish has joined #sandstorm
FredFredFred has joined #sandstorm
FredFredFred_ has quit [Ping timeout: 240 seconds]
FredFredFred has quit [Ping timeout: 248 seconds]
pie_ has joined #sandstorm
samba_ has joined #sandstorm
jemc has quit [Ping timeout: 260 seconds]
samba_ has quit [Ping timeout: 264 seconds]
afuentes has joined #sandstorm
digitalcircuit has quit [Ping timeout: 240 seconds]
digitalcircuit has joined #sandstorm
digitalcircuit has quit [Ping timeout: 255 seconds]
digitalcircuit has joined #sandstorm
bodisiw has joined #sandstorm
gambatte has joined #sandstorm
wolcen has quit [Ping timeout: 255 seconds]
jemc has joined #sandstorm
wolcen has joined #sandstorm
pie_ has quit [Ping timeout: 245 seconds]
<frew>
hey guys, I have a git repo in sandstorm and pushing to it is timing out
<frew>
it looks like the credentials changed in the past few months?
<frew>
is this something known and should I be worried?
<dwrensha>
frew: did your local ~/.git-credentials file get cleared somehow?
<frew>
hmmm..
<frew>
I don't think so?
<dwrensha>
hm. I suppose that if it were a credentials problem then I would expect a 403 error, not a timeout.
dyn has joined #sandstorm
<dyn>
Hi all, longtime supporter/fan of sandstorm, but just now getting around to setting it up myself on a new box and I'm having some trouble. I installed it fine on 16.04 LTS, but a few things aren't working.
<dyn>
For whatever reason, the sandstorm server isn't using 6080 for apps or static resources
<dyn>
and setting HTTP_PORT in the config to another port doesn't change that
<dyn>
I still have links going to "static.local.sandstorm.io" port 80, which isn't right
<dyn>
I should point out that I'm using the *.local.sandstorm.io WILDCARD for now
<dyn>
I guess my question is, why does the server generate links pointing to port 80, if the server doesn't listen on port 80 (which I confirmed with netstat)
<dyn>
Same thing for launching apps: "Request URL:http://97b5ec7566c7277d2eb62c2a8bdda2dc.local.sandstorm.io/_sandstorm-init?sessionid=195e21c4b6a5e1664a7820a9e2218accd5ec471a464f32a62948376b84c70279&path=/"
<dyn>
it seems like something isn't respecting HTTP_PORT or somehow not defaulting to 6080 anymore? I don't have any error logs about it not binding to port 80 and I installed using the script w/ all recommendations
<dyn>
nevermind... I fixed it
<dyn>
WILDCARD_HOST needed to have ":6080" on the end
<dyn>
this maybe should be more clear? *shrug*. Thanks for sandstorm! ttyl
<Guest3349>
My sandstorm server hardware (not the HD though) recently died and i do not have a recent backup (personal system, not corp). I was hoping to get some pointers on how to recover the contents of that sandstorm instance.
<Guest3349>
I was considering installing FUSE on my Mac so I could read the ext3 filesystem (Ubuntu 16.04 install on sandstorm drive)
<Guest3349>
I also have an available desktop. Could I install the drive in that PC and boot Ubuntu? Would that work?
afuentes has quit [Ping timeout: 240 seconds]
<mokomull>
Guest3349: That's a more generic question than Sandstorm, but as a first stab at it, yeah. Linux tends to be a lot more stable than, say, Windows when you move the filesystem between computers.
funwhilelost has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
neynah has joined #sandstorm
jemc has quit [Ping timeout: 260 seconds]
<Guest3349>
yeah very general but I know someone here would know. I will move the drive to another system and then to a proper backup according to the docs.