« first day (1815 days earlier)      last day (3202 days later) » 
00:00 - 23:0023:00 - 00:00

11:00 PM
@Bob No worries at all
 
Bob
At this point the goal is to figure out which file, if possible
Though perhaps you could do a repair install
 
I couldn't see anything obvious either in windbg or in BlueScreenView
I'll pastebin each
 
Bob
Sometimes you need to dig into the error params
Though this stinks crc fail or similar
In which case it's not guaranteed that the filename is even retrievable
And would likely need a kernel dump at least
stinks of*
 
@DragonLord Well to be fair the gcc error does tell you exactly what's wrong, but in a manner less friendly to newer developers
 
@Bob Doesn't sound great, but we'll see how we go
pastebin.com/D5qsiYnn (BlueScreenView)
 
Bob
11:04 PM
The clang error is arguably too noisy for more experienced devs
@bertieb I'll take a look when I get to a computer
 
@Bob Cheers :)
Briefly, stack text:
[...] nt!KeBugCheckEx
[...] nt!PopGracefulShutdown+0x2c9
[...] nt! ?? ::OKHAJAOM::`string'+0xe30
BugCheck C0000102, {ffffc00001003150, 0, 0, 0}

Probably caused by : ntkrnlmp.exe ( nt! ?? ::OKHAJAOM::`string'+e30 )
 
@Bob Would have to agree with that to an extent; pretty sure there's a way to suppress that
 
Bob
11:32 PM
@bertieb ooookay
You might get some more info with a kernel dump and !object with the first arg (ffffc00001003150 this time)
or you can try that on the minidump
 
@Bob I'll fire up safe mode and switch the dump type
 
Bob
Oh, another thing you could try: selectively enable individual drivers until you find the breaking one :P
 
1: kd> !object ffffc00001003150
GetUlongFromAddress: unable to read from fffff8013dfce010
Could not read ObjectType address
 
Bob
@bertieb also, try inspecting the memory at that address: msdn.microsoft.com/en-us/library/windows/hardware/…
I'm hoping it contains the filename
(there's not much documentation on the ntfs errors, unlike the normal bugcheck errors)
@bertieb could also try ds on that address
 
!
 
11:36 PM
-1
Q: Is there an ethical standard for using Google Analytics?

1.21 gigawattsI'm setting Google Analytics for my website for software that I'm about to release. When signing up there were numerous options and all were labeled secure and does not share data but some of it was confusing. Is there a set of options that are best for privacy of the user but also helps me as a ...

<sigh>
 
Reference to syngery (km)
@Bob At that address I mean
 
Bob
o.O
was that with ds?
try uninstalling synergy from safe mode :P
actually, don't
make a full backup first
 
ds just shows a bunch of what I assume is unprintables
 
Bob
in case something else breaks
@bertieb then how'd you find it?
 
I could disable synergy service from autostarting
 
Bob
11:38 PM
@bertieb won't disable the driver
 
@Bob memory browser as you suggested
 
Bob
or maybe it will, who knows
 
Hmm, I guess
 
Bob
@bertieb ah
 
Bob
11:39 PM
ds expects a certain string format
 
I really wish that case wasn't $500 :-(
 
Bob
not null-terminated
@bertieb won't hurt to try disabling the service
 
@Bob Just doing so, will reboot
 
Bob
fingers crossed!
 
The reference might be to a file is the confusing thing
 
Bob
11:41 PM
@bertieb ?
 
Memory location reads \apps\synergy
 
Bob
hm
 
That's about it
 
Bob
odd
 
Oh hello
D:\apps\syngergy:

The file or directory is corrupted and unreadable
Miiight be the culprit
 
Bob
11:42 PM
ha, that would do it :P
noooowww... how to fix it? :P
 
Good question
 
Bob
hm. could try deleting it and seeing what happens
you might get a missing file error, but not an ntfs error
 
How is Windows/NTFS at handling the removal of corrupted directories?
 
Bob
backup first
 
shrug it renames okay
So will give that a try and see what happens; can always rename it back
 
Bob
11:43 PM
@bertieb the name is part of the parent dir's metadata
the object the dir itself points to is broken
 
@Bob Yeah, but anything that references it should get a os file not found when trying to do a read I'm hoping
As opposed to a OHGODSCORRUPTION error :P
 
Bob
that's a git commit tree :P
(from the git project itself)
@bertieb yea, hopefully
I suspect Windows will cope with a missing non-critical driver
but just suicides at the first sign of corruption during driver load
 
@Bob looks like preparing for a release
 
If this is the problem, the more worrying question is how did it become corrupted in the first place?
 
(I am assuming those are pull requests merged)
 
11:47 PM
Success! (provisionally)
 
Bob
@bertieb you really need to check out the HDD SMART data sometime
and run 5+ memtest passes overnight
 
@Bob I owe you a postcard (as mentioned) / coffee / beer at the very least
I could swear windows monitors SMART these days
I'll pull it up under Linux otherwise
 
Bob
@bertieb don't think so, but there's quite a few tools to do it (e.g. crystaldiskinfo)
 
I'm sure I've seen notification ballon SMART warnings before
An older mobo I have gives SMART error notifications (that you have to F1 to dismiss before booting) during POST
 
Bob
hm. maybe just on fail states?
a couple reallocations isn't necessarily a failure yet, but could be indicative of one soon
 
11:50 PM
Will smartctl just now anyways
@Bob Ah, true that, good point
0 reallocated sectors
Doing a short self-test --> passed
 
Bob
hm
odd :\
definitely make a backup if you can
 
Could have been a very unluckily mis-timed write
Yeah
At this point, I'm very happy that it didn't need a nuking
Couldn't have done it without your help @Bob
 
Bob
@bertieb ah, just a couple of tips and a wild goose chase (bootloader) here and there :P
you did most of it yourself
 
So pick your just reward from the ones mentioned before, or I can write this all up as a question, and you can get the rep for the answer
I did the goose chasing in fairness (came in asking about bcd etc), you had the crucial windbg info
 
Bob
@bertieb if you can make it suitably generic (about the ntfs corruption error, or that specific error code), then that could help - there was a severe lack of resources for that code :\
 
11:59 PM
I should write the question up anyway
There's yeah
You beat me to that :P
 
Bob
though it might be better if you self-answer - you know the exact steps you took
 
00:00 - 23:0023:00 - 00:00

« first day (1815 days earlier)      last day (3202 days later) »