« first day (1298 days earlier)      last day (4018 days later) » 

Bob
Bob
17:00
hi
Morning all
how's life?
going good
computer's giving some problems - random freezes
not bsods, just freezes & stops responding
Bob
Bob
@Sathya primary suspect: drivers
likely drivers: video
17:02
any tips on getting to the root of the problem?
hm but haven't updated video drivers in a while
Bob
Bob
suggested fix: try updating video drivers (uninstall/reboot/loop until they're all gone first)
hm ok
Bob
Bob
@Sathya other possibility: overheating
not overheating, temps are below 40s
Bob
Bob
eh... I'm out of ideas
or we can always go back to the basics
1. how long has this been happening
2. what was the last thing you did
2a. if the answer to the last question was nothing, what was the last thing you did before nothing (I love asking that)
other random ideas: try a livecd
17:04
:)
Bob
Bob
3. how often does this happen?
4. does it resume after a bit or stay not responding?
no discernable pattern
could also be ATA transfer failures during reads or writes
(HDD issues...)
happens when idle
when you freeze, can you move the mouse?
17:05
happens when under activity
Bob
Bob
@Sathya once a day, 5x, one a week, that kinda thing
@allquixotic nope, keyb stops responding
wait, windows or linux?
Bob
Bob
@allquixotic oh yea, forgot that one
Win 8.1
I suspect HDD, but SMART don't say anything's bad
another suspect is networx
Bob
Bob
17:06
@Sathya you can always try a benchmark/stress test tool (backups first!)
or swap out the SATA cable, sometimes that dies
@Bob I'll same maybe once every other day.. but I've had 7+ day uptime & then 3 freezesa day
Bob
Bob
yay! more random!
...
I like giving you guys a tough time eh
Bob
Bob
incidentally, I recently fixed a similar issue with a video driver update (which is why i suggested that :P)
also appeared out of the blue
AMD?
Bob
Bob
17:08
nvidia
@Sathya you never did answer - does it recover after a while?
@Bob I think not, never waited long enough ( max~ 2-3 minutes & then I hit reset)
Bob
Bob
also, do you happen to use a SSD? there was this weird bug that was causing similar freezes from some command... there's a registry edit to fix that
but I can never find it in a search
it's bookmarked on the work pc :\
but there's no response from keyboard (ie lights don't go on/off
@Bob no ssd, a HDD
Bob
Bob
@Sathya try waiting 5-15
no response from kbd lights == kernel no longer handling interrupts
it's a BSOD that's so D it can't even render a :(
3
Bob
Bob
17:11
@allquixotic ditto for lack of cursor response
but it is recoverable, sometimes
I think not
Bob
Bob
usually when the blocking operation hits a timeout
@allquixotic I've had the same situation many times, often recovering
with nvidia drivers, it hits the user mode display driver timeout and restarts it
unless it's blocking in the main loop, in which case there's no other thread to rescue it and trigger the timeout
you can't timeout synchronously with only one thread of execution
Bob
Bob
with the weird SSD SATA bug, it hits some IO timeout and resumes normal operation
if the core spinlock is interrupted, you're dead
17:12
once a blue moon, I hit DPC_WATCHDOG_VIOLATION from networx ( ~ once very other week) but this freeze is prevalent
i get those too, even rarer though
the freeze hit me today middle of a hearthstone game -.-
what gpu?
AMD 7750
:O\
i have a 7970.. maybe we're twins, having the same issue with AMD driver :D
since i get DPC_WATCHDOG_VIOLATION too
Bob
Bob
17:14
oh yea
Bob
Bob
had similar issues some time ago when my ODD had a bad SATA cable
thing is, with that kind of freeze, the behavior is non-deterministic, so there's no guarantee you'll see the BSOD
the remaining freezes are probably the same root cause but so hosed it can't even display the BSOD
Bob
Bob
@allquixotic eh, I've found freezes are more often waiting for some driver to finish an operation while BSoDs are more the driver does something obviously illegal
I say "I've found", but that's what's supposed to happen anyway
and I can say it does happen in practice
which is also why some types of freezes will recover
17:15
@Sathya also, hearthstone? your soul shall be mine!
we need to play sometime
watch. your. back.
sathyabhat#6507
(I suck at it)
@Sathya the light shall burn you!
Bob
Bob
again - try updating drivers (full uninstall/restart/loop + update)
try swapping out the SATA cables if possible
or at least replugging them
17:16
okay
Bob
Bob
try a HDD stress test
wilco
Bob
Bob
annnd I'm gonna sleep
@Bob which tool? HD Tune?
if all else fails, give up and hope that your hearthstone freezes again when you're in a game with me, for a free win
Bob
Bob
17:17
@Sathya should do
anything that hammers the HDD with requests
but make sure you have a backup!
@allquixotic ha. will be my first win :D
@Bob yeah, that I do
@Bob Thanks for mentioning the StackOverflowException stuff earlier.
I just got a support ticket about the .Net runtime crashing without executing any of the catch/finally handlers or logging errors
I swear this stuff could not be more perfectly timed
@DarthAndroid lol awesome
that whole conversation is my fault though
so you should thank me
Thank you!
Bob
Bob
:P
@DarthAndroid there's also OutOfMemoryException that can do this
17:21
It's always fun learning new stuff :)
Bob
Bob
now the problem is trying to track down the cause...
without logging! joy!
@Bob I'm fairly confident it's not an OOM in this case, but I suppose that's possible.
amen!
i'm really surprised .NET doesn't have Errors like Java for uncatchable system/environment problems
@Bob Not my problem :D
17:21
OutOfMemoryError >>>> OutOfMemoryException
you can't catch OOM and do anything meaningful
Bob
Bob
@allquixotic I guess at least the MS CLR is too close to the system for that.
I get the impression the JVM has a little more abstraction for things like stack allocation. Could be wrong (never really looked into it).
It does
it gets hairy with the JIT
Mono and presumably .NET JIT too
Well, yeah, but I already know .Net is using the native thread stack
Bob
Bob
Pretty sure it's the system terminating the entire runtime.
Which raises the question why they didn't do something like @DarthAndroid suggested with a larger stack guard.
17:24
Can a process set custom stack sizes per-thread?
Bob
Bob
@DarthAndroid yep
or does it require system privileges to do that?
(well, to make them huge)
Because I've had to give 5MB stacks to some java apps that were handling large data
Bob
Bob
You can also create a guard page with Win32.
@DarthAndroid No, you can do it with System.Threading.Thread
or, if you want the main thread to have a different stack size you can edit the PE headers
some tool in the Windows SDK, I think
.NET's default stack size is actually quite big
I remember I was looking into reducing it at one point
was spinning off a thousand threads to stress test a java message queue (which was actually about 3k because of the stupid JMS client spinning off two of its own and blocking - it didn't do async requests either)
Then Java might have native stack frames for the java stack frames
@DarthAndroid would throwing things into arrays of integral types (nested in classes as necessary) make it allocate them on the heap instead? (performance--, but, no stack size explosion, which could end up being a net win)
arrays implicitly "extend" Object in Java, IIRC
so they'd be heap allocated
17:38
'Pick you up at eight?' 'Nine.  I've got to re-mine the driveway.'
2
@HackToHell O_o
18:00
@allquixotic Arrays are definitely heap-allocated
The only things that get to use the stack are primitives and pointers to non-primitives.
18:26
@DarthAndroid I guess maybe your app was so complicated that it literally got "pointered" to death -- enough to need 5 MB of stack -- pretty impressive
or they just had a ton of long primitives in a long call chain, or recursion or such
Oh, I'm sure it was recursion. You'd have to have so many functional parameters and individual variable declarations to overflow the stack that I don't think I've ever seen that happen as a leading cause.
(I was using a profiler and had left it running doing CPU sampling for >24 hours
it does some recursive copy when you try to snapshot it, I'm pretty sure)
18:45
@DarthAndroid it's a little easier to blow the stack in C++, though, I find... frameworks like Qt actively encourage you to stack-allocate your data structures, so you can have 2d arrays of pixels and other such "large" data on the stack sometimes
in Qt, strings still get pooled into the heap (COW) and the stack-allocated bit is just a reference (IIRC), but much other stuff has the actual data payload on the stack
19:04
@allquixotic Oh, yes, but you don't have that in java, and the profiler is written in java
well, to be more accurate, the profiler is the JVM, this is just a client that connects
Bob
Bob
19:40
@allquixotic I kinda get the feeling they're not the type to fall for those :P
@Bob Are you in Sydney?
Haven't seen one of these by chance?
Bob
Bob
@MichaelFrank Nope. Where are they?
I haven't been to the CBD since last Friday.
Mardi Gras is next weekend, I think, which might explain that.
Yeah, there's about 10 of them around apparently.
Bob
Bob
o.O
TIL the Sydney Mardi Gras isn't quite like one in other places...
@MichaelFrank You got a list of locations?
Though, that looks like an ANZ one (can barely read the card image). I'm with CommBank - wouldn't have much reason to use those.
@Bob Yea, they are ANZ.
Bob
Bob
@MichaelFrank Well, I'm heading out today. Anywhere I should take a look at?
SUMMARY: During normal operation or in Safe mode, your computer may play "Fur Elise" or "It's a Small, Small World" seemingly at random. This is an indication sent to the PC speaker from the computer's BIOS that the CPU fan is failing or has failed, or that the power supply voltages have drifted out of tolerance. This is a design feature of a detection circuit and system BIOSes developed by Award/Unicore from 1997 on.
Bob
Bob
@Hennes wut.
Wut ^ 2
I just checked the date. It is not the first of April.
Bob
Bob
20:02
@Bob •ANZ’s GAYTMs can be found in various Sydney locations on Oxford Street, Pitt Street, George Street, Castlereagh Street, Surry Hills and Bondi Junction
Bob
Bob
@MichaelFrank Damn. I'm not passing those today.
Maybe Friday.
Wait, Castlereagh? O.O
Problem is, I have no idea where ANZ puts their ATMs...
couple of Westpac, CommBank...
Abby T. Miller on February 24, 2014

Welcome to Stack Exchange Podcast #55, recorded on Friday Thursday the 13th with your hosts Joel Spolsky, David Fullerton, and Jay Hanlon! Today’s episode is brought to you by the city of Sochi, Russia.

It’s been a long time since we last recorded, so we have a lot to talk about, and we’re going to skip most of it. First we’re going to talk about all our brand new sites, so Joel learn about them for the first time. Pets is a site for (you guessed it) pet owners to wonder why their cats like to watch them making the bed. Also, we already talked about this site. Moving on …

@Bob I'm sure you'll find it difficult to miss them now. :P
Bob
Bob
True enough!
WUT
> The Nokia X is a 4-inch phone that will cost 89 Euros, and is available immediately. The Nokia X+ will be 99 Euros, and it will have more storage. Nokia is also rolling out an XL phone, which has a 5-inch screen. The X+ and XL will be out in early Q2. The X phones will be rolling out “broadly,” but starting in “high-growth markets.”
holy shit
I... actually kinda want one
> Nokia says its primary focus is the Lumia line, which is based on the Windows Phone platform. Nokia X is a lower-end product going after developing markets.
or not :(
20:14
!!tell 13953076 no
Bob
Bob
@allquixotic :(
also, the Nokia X / X+ is an AOSP non-Gapps phone.... "Nokia Store" lolol
Bob
Bob
@allquixotic meh. root => fun
root = no Verizon. lol
Bob
Bob
20:15
and, really, I love Nokia hardware
@allquixotic I don't live in the US :P
@Bob this isn't your father's Nokia (or something like that)
Bob
Bob
we have these things called SIM cards
basically the hardware is going to suuuuuuuck
> lower-end
Bob
Bob
@allquixotic ...I have a L520 I rather like
don't expect this to be a brick like old Nokia devices
Bob
Bob
20:16
..
I had an E65 (two, actually), N8 and Lumia 520.
it probably has a plastic screen :D
Bob
Bob
None of them are bricks
Nokia doesn't know what it wants to do or who it wants to side with. it never has
Bob
Bob
@allquixotic gah. Maemo! D:
it's been all over the freaking board
Bob
Bob
20:17
Symbian^3 was actually halfway decent
Maemo, Android, Microsoft, Symbian, Qt, ... I mean...
Bob
Bob
had it in lieu of Android around that time
hardware was lower-power, but it worked well
and Maemo had a compat layer for Android apps
...but nooo, they dropped it
it's really embarrassing how Nokia has been kind of bouncing between various platforms in the past few years... they should just pick a side and stick with it, grow some internal expertise with that platform and champion it
Bob
Bob
@allquixotic Still, I like their hardware.
And I'm talking current-gen.
@Bob lol, PalmSource/ACCESS's ACCESS Linux Platform (based on the idea of porting desktop GTK2 apps to the phone "seamlessly", with multi-touch support, using a lot of runtime stuff) had a Garnet VM for PalmOS 5.x
another albatross that never flew
@MichaelFrank kinetic energy at work
Bob
Bob
@allquixotic So... work at work?
@Bob -_- SRIV 60% off on HIB right now
@Bob yuss!
Bob
Bob
@allquixotic .....
how much was the steam sale discounted?
i dont remember T_T
Bob
Bob
20:21
-_-
25AUD lowest here
without the DLC
you sent me 41 usd for all of the SRIV content
Bob
Bob
20USD lowest
again, without DLC
@allquixotic Full cart including HIB DLCs is 44 USD
But there's been a couple of DLC packs released in the last month.
So, about the same.
Bob
Bob
@allquixotic I guess the advantage (for me) of HIB is I can pay the US price directly :P
except the base game costs as much as the AUD price on sale anyway, and even more after conversion..
Bob
Bob
20:32
@MichaelFrank Humble Indie Bundle
though it's really more HB these days
@Bob HAAAB? :D
Are you sure you're not confusing the Weekly bundle and the regular bundles? :P
Ugh... the conversation is so much better here than over at the Gaming.SE chat room!
20:48
@MichaelFrank what's wrong with The Bridge?
aside from...... eh, nevermind
@allquixotic They were ripping into this guy because he friend requested some girl.
Is there a good program for parsing Apache log files?
ParselTongue
@MichaelFrank Was that to me?
Oh yea, sorry.
21:01
@MichaelFrank heh...
I searched for that, but don't see any programs
@CanadianLuke it was a joke. :X
Bob
Bob
@CanadianLuke vim
well, that's what I use anyway
how exactly do you parse a log file, apart from some grep and vim?
@Bob he's probably still trying to work out how to parse Apache logs using the magical snake language. :D
Softwarerecs.SE.com? Maybe?
@Bob I was hoping for a tool that I could feed my log file into, and it could summarize what is in there
Bob
Bob
21:13
"summarise"?
"you have 5000 errors"?
Really not sure what you can summarise from an Apache log...
"The chamber of secrets has opened!"
Ha ha, I got a program called "Apache Log Viewer". Free for basic use, which is all I want
21:38
@CanadianLuke and what does it do, just display the literal text of each Apache line in a table?
2
22:04
And filter, which is useful
Bob
Bob
...that's what grep is for
@CanadianLuke I have to admit, that's a little silly from the perspective of someone who knows how to use basic UNIX tools
22:19
@allquixotic sometimes you just want buttons to click.
@Bob I understand I can use Grep, but as @MichaelFrank states, sometimes you just want to push buttons!
My actual goal is to see what requests aren't going to where my site is located, and block those IP addresses
Bob
Bob
@CanadianLuke That's what all the grep GUIs are for
The extra legwork of pulling the logfile off of a server to my local system negates any desire to click buttons for me :X
Also, grep -v (your site) | grep -e '[[:digit:]]{1,3}\.[[:digit:]]{1,3}\.[[:digit:]]{1,3}\.[[:digit:]]{1,3}'
add -o to the second grep if you want a list of just the IP addresses that you can plug into another pipe
@DarthAndroid Nope
Huh, this is really cool
@CanadianLuke Yup, though I botched some of the options
Should be grep -E or egrep for the second, and the first grep needs a filename
but
sudo grep -v "dragonzone.net" access.log | grep -v "10.0.3.2" | egrep -o '[[:digit:]]{1,3}\.[[:digit:]]{1,3}\.[[:digit:]]{1,3}\.[[:digit:]]{1,3}' | sort | uniq
Totally works
(The extra grep -v 10.0.3.2 is to filter out a local network that accesses by IP address)
A lot of these are google bots I think
Wtf baidu is spidering my site
._.
64.15.159.21 - - [24/Feb/2014:10:35:47 -0500] "GET /w00tw00t.at.blackhats.romanian.anti-sec:) HTTP/1.1" 200 396 "-" "ZmEu"
Whelp, guess I'm already being pen-tested.
I suppose I'll lock things down when I get home.
22:35
Welcome to nginx!
If you see this page, the nginx web server is successfully installed and working. Further configuration is required.

For online documentation and support please refer to nginx.org.
Commercial support is available at nginx.com.

Thank you for using nginx.
Indeed :)
Need to shut that off.
Only using subdomains currently
@DarthAndroid So what does that tell you?
What does what tell me?
The output of the full command, the line about being pentested, or the default nginx page that you pasted to me?
The line I replied to, which is the output I guess.
It tells me that you visted dragonzone.net, or typed in the IP directly.
Mostly because I'm not using that domain yet, but I'm using several subdomains off of it.
and I haven't locked down the root page
22:46
12 mins ago, by Darth Android
64.15.159.21 - - [24/Feb/2014:10:35:47 -0500] "GET /w00tw00t.at.blackhats.romanian.anti-sec:) HTTP/1.1" 200 396 "-" "ZmEu"
What does that tell you.
That in particular tells me that there are romanians probing my server for vulnerabilities
It is preceeded/proceeded by attempts to open two common MySQL admin panels
(MyAdmin and phpMyAdmin)
This is what I mean: pastebin.com/PpsbdUSd
Can't see pastebin at work. :/
115.238.101.45 - - [23/Feb/2014:09:14:37 -0500] "GET /phpmyadmin/scripts/setup.php HTTP/1.1" 200 396 "-" "ZmEu"
115.238.101.45 - - [23/Feb/2014:09:14:37 -0500] "GET /phpMyAdmin/scripts/setup.php HTTP/1.1" 200 396 "-" "ZmEu"
115.238.101.45 - - [23/Feb/2014:09:14:37 -0500] "GET /w00tw00t.at.blackhats.romanian.anti-sec:) HTTP/1.1" 200 396 "-" "ZmEu"
115.238.101.45 - - [23/Feb/2014:09:14:37 -0500] "GET /MyAdmin/scripts/setup.php HTTP/1.1" 200 396 "-" "ZmEu"
115.238.101.45 - - [23/Feb/2014:09:14:37 -0500] "GET /pma/scripts/setup.php HTTP/1.1" 200 396 "-" "ZmEu"
23:09
Ahh okay, I see. I was just curious what the number and string near the end were.
well
200 is the status code (e.g., 404 = file not found, 200 = ok)
not sure what 396 is
"-" is the requested subdomain
or the referrer
one of the two
the last string is the user agent
@DarthAndroid Ahhh that worked! Now, I just need to see what they're accessing, or if they're bots
With nginx, it's easy to check the user agent
@DarthAndroid I use Fail2ban to block these accesses
65.55.213.69 - - [23/Feb/2014:08:01:50 -0500] "GET / HTTP/1.1" 200 396 "-" "msnbot/2.0b (+search.msn.com/msnbot.htm)";
220.181.108.146 - - [23/Feb/2014:08:27:46 -0500] "GET / HTTP/1.1" 200 612 "-" "Mozilla/5.0 (compatible; Baiduspider/2.0; +baidu.com/search/spider.html)";
23:15
The ZmEu user agent string is blocked on sight
@DragonLord I started reading about Fail2ban, but I need time to sit down and iron out what I want
@DragonLord Yeah, need to look into that
My Fail2ban configuration is available upon request
@CanadianLuke There's an example of Bing and Baidu's spiders
@DarthAndroid These accesses are normal--don't block them, or you won't be found on search engine result pages
23:16
@DragonLord I know
robots.txt is how you tell those to stay away :)
Guess who's got internet at home starting today!
And guess who's having the privilege to pay for 5mbps to get 2,5mbps
Oh, and the router provided by the ISP has a 2nd, unstoppable, unconfigurable SSID sharing my home connection to the world.
Are you at least getting free service?
(The IS tech, obviously, never mentioned it and I had to google for the symptoms)
46 secs ago, by That Brazilian Guy
And guess who's having the privilege to pay for 5mbps to get 2,5mbps
Whoops, missed that
I have a bit of a concern with one of my blog posts
23:18
Emphasis on the pay for
I suppose calling up support is out of the question?
A blog post about a responsive redesign of my main website is attracting a lot of spam attempts, and my server logs indicate a lot of rather unnatural traffic to that blog post
or otherwise getting the SSID turned off?
It's not really much more than a meta post on my website--should I delete the post?
@DarthAndroid Not exactly, it's just they're utterly useless.
I managed to turn it off in the end, thanks to a random forum post
23:20
Oh, good.
I just had to access 192.168.1.1/undocumented-url-never-mentioned-anywhere
Hits	%	Bandwidth	URL
674	8.07%	15.42 MiB	/index.php?/archives/8-Main-site-redesigned-for-responsiveness.html
462	5.53%	10.34 MiB	/index.php?/archives/3-Issues-in-the-blog-software-resolved-Updated.html
346	4.15%	10.52 MiB	/index.php?/archives/5-Dhrystone-Running-a-25-year-old-benchmark-on-modern-hardware.html
306	3.67%	6.72 MiB	/index.php?/archives/6-IPv6-enabled.html
@ThatBrazilianGuy ask & answer on SU.
But seriously, WTF? Sneakly inserting a 2nd SSID on my router? Not even malware creators are so bold and evil.
That's actually a thing, though not quite on those terms
Usually it's <public Wifi Hotspot provider> offers you a free router or free router/connection
23:23
Three of the four posts listed above are garnering are getting amounts of traffic that are unexpected for content that is about the blog or the website itself
@DarthAndroid It's a Brazilian ISP. All the google searches will be in Portuguese, there'll be barely no hits and no users helped.
@CanadianLuke, should I delete every one of these posts except the Dhrystone one?
@ThatBrazilianGuy Ah, blog post it is, then. :)
@DarthAndroid Yes, apparently my paid carrier thinks it is a great idea to do that without telling its clients
They seem to be attracting unnatural spambot traffic
23:24
@DarthAndroid I don't have a blog
I lack the diligence and focus to have one :(
@DragonLord not sure what to tell you.
@DragonLord Insert a block rule on .htaccess or something like that
A close read of the server logs indicate that a large portion of accesses to the meta posts are from spambots trying to post spam comments (they're automatically blocked)
@DragonLord Nah, just rename it
shrugs
As to why it's those posts, I'd say probably just because the bots found links to those posts and not the other ones
23:27
@CanadianLuke: I think I'm going to close comments on those posts. That will stop the spam accesses.
If you're wondering about the how the table was generated, I used GoAccess
Bob
Bob
@that simple, wrap it in tinfoil and get your own Wi-Fi router
Bob
Bob
@Darth the tinfoil approach is more certain and tinfoilly
23:41
GoAccess works for me!
> Note: Due to excessive spamming attempts on this post, comments have been closed.
@CanadianLuke It's amazing how much information server logs contain...
I agree
I apparently need a favicon though
@ThatBrazilianGuy: hmm, second SSID. You could set up a second, stronger AP with the same name
then most systems would 'prefer' the stronger AP, connect to that, and find themselves seeing whatever you want them to see ;p

« first day (1298 days earlier)      last day (4018 days later) »