-
Notifications
You must be signed in to change notification settings - Fork 95
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
Extremely random crashes caused by unknown error #150
Comments
No, streamer isn't causing this. Your script is. |
I don't think you understand the slightest about debugging sir, with all due respect, if it were a script error, the issue would've been printed in the debug log, and I wouldn't be here writing this issue right now. So if you aren't the developer, please don't give your input. |
I believe this is a buffer overflow within the plugin, which is why it's been posted here. Our system has a map changing routine which destroys and recreates map objects on demand. Thus, one map is destroyed and another is recreated for the player. Crayder it would be a real honor too if you contributed only useful information to this issue. Good luck. |
[11:55:23] [debug] Server crashed due to an unknown error Compiled streamer in debug mode, and we got this. |
Could you run the server in GDB? |
I got the candy you requested (: backtrace: registers: current instructions: threads backtrace: Thread 8 (Thread 0xf2e6bb70 (LWP 14321)): Thread 7 (Thread 0xf386cb70 (LWP 14320)): Thread 6 (Thread 0xf426db70 (LWP 14319)): Thread 5 (Thread 0xf4c6eb70 (LWP 14318)): Thread 4 (Thread 0xf7000b70 (LWP 14317)): Thread 3 (Thread 0xf7de8b70 (LWP 14316)): Thread 1 (Thread 0xf7ff28e0 (LWP 14312)): |
And if it's of any assistance here is how it can be replicated: 1.) Player has map 1 streamed in for him. |
I actually commented on the wrong issue when I said it was the script, but now with that last comment said I think this might also be the script. |
Again, Crayder I would be very much indebted if you posted a response based in knowledge; instead of thoughtless utters. This is a broad issue - maps are streamed in for a player, he goes AFK and during which time objects are destroyed. He returns and the plugin crashes. Please read before posting. Thanks. |
It could be the script, but it's odd how Can you also reproduce this in an empty script with a few objects being created and destroyed? |
I did not compile the latest streamer plugin in debug mode, I was running on no sleep during the time I ran that instance of our sa-mp server in gdb. However i've compiled it in debug mode this time and will be providing you with the results from the next run. Will also probably try an empty script with object creation & deletion to see if we can replicate it outside of our code. |
Alright so after compiling 2.8.2 in debug, this is what I got for you, some lines now. backtrace: registers: current instructions: threads backtrace: Thread 8 (Thread 0xf2e6bb70 (LWP 17475)): Thread 7 (Thread 0xf386cb70 (LWP 17474)): Thread 6 (Thread 0xf426db70 (LWP 17473)): Thread 5 (Thread 0xf4c6eb70 (LWP 17472)): Thread 4 (Thread 0xf7000b70 (LWP 17471)): Thread 3 (Thread 0xf7de8b70 (LWP 17470)): Thread 1 (Thread 0xf7ff28e0 (LWP 17466)): |
Thanks, that's very helpful. I can see what's causing this now. Actually, I'm rather shocked it hasn't been caught before... |
If an item currently visible for a player has been reassigned to the global cell in between updates, it could cause a crash. Related to #150.
All the while we were convinced that the objects creating/destroying was causing this, but I discovered yesterday evening that it was in fact zones. Thanks Incognito, we will test and report back! |
The crashes are still occurring, what's odd is that they're always the same, will post a new gdb within the upcoming hours. |
Additionally, data manipulation natives such as Streamer_AppendArrayData, Streamer_SetFloatData, Streamer_RemoveArrayData are claiming invalid data is specified. Sometimes it works, sometimes it doesn't in light of the aforementioned commit. |
GDB also fails to give any information now related to the crash, which is also very confusing. |
@TheDarknessz0rz @SFSEdev Same for me, it's still crashing, and after recompile those functions are claiming that invalid data is specified. So i was forced to get back on 2.8.2 official release and to delete area system. |
You'll need to compile with the newest streamer.inc from here. |
Didn't cross my mind, how embarrassing... Nevertheless, it is still crashing and we will provide as much information as possible in the upcoming hours. Thanks! |
The crash reported in this issue thread has been resolved, we are currently tracking another crash unrelated to streamer, this can be tagged as fixed. Thank you for the speedy replies & fix incognito! |
Share your solution so others can use it when/if they have this issue. Also tell us how to reproduce it. |
Did you even bother reading the replies above? (psst... this isn't the first time your inability to read has shown) |
Mind reading my comment again? Neither of you posted the fucking fix or how to reproduce the issue (NOT the same as how you came across the issue, we need to know WHAT you found causes the fucking issue). These are key ingredients to helping others in the future.
My inability to read what exactly? I read the crashdetect information and it is something that can easily be reproduced (and I've come across this issue many times and still do) WITHIN THE SCRIPT. |
Crayder I don't think you understand, the issue we've posted here was in streamer code, which Incognito has pushed a fix for. If you did read the issue thread, you'd see that incognito pushed a commit that solved the issue we opened here...... How would it be possible for us to speak about a fix for a piece of broken code that wasn't ours to begin with? Incognito has outlined the issue himself & patched it. Are you sure you've actually read it? Just so things do not divulge further off topic, @samp-incognito you may tag this issue as fixed, your patch commit has fixed the issue. The crash issue we experience now appears to lie deeper in the sa-mp code, as our method of invocation is not preferred by the sa-mp natives. |
Oh really? So I'm supposed to just assume that his push actually did fix the issue even though right after he pushed that you said the following?
Yeah, sorry bud but I don't think so. If that WAS the fix and it DID work, you should've clarified that when you said this... which btw you said way after you said what I quoted above:
|
Full of spite, looking to shun others work and then ask for their solution, you're just full of heat, I posted a day ago that the issue reported in this thread was fixed, and now have ran into a separate issue. you just don't learn... Also, this will be the last time I reply here, we are not entitled to explain anything to you that you can't already learn from reading this thread. Thank you incognito for the help you've provided us, sorry to see that this thread has turned into a fuckfest of nonsense..... wasn't our intentions. |
[22:18:55] [debug] Server crashed due to an unknown error
[22:18:55] [debug] Native backtrace:
[22:18:55] [debug] #0 0058b55b in _ZN10StackTraceC2EPv () from plugins/crashdetect.so
[22:18:55] [debug] #1 005845df in _ZN11CrashDetect20PrintNativeBacktraceERSoPv () from plugins/crashdetect.so
[22:18:55] [debug] #2 005857a2 in _ZN11CrashDetect20PrintNativeBacktraceEPv () from plugins/crashdetect.so
[22:18:55] [debug] #3 00585bf6 in _ZN11CrashDetect11OnExceptionEPv () from plugins/crashdetect.so
[22:18:55] [debug] #4 0058b2fc in ?? () from plugins/crashdetect.so
[22:18:55] [debug] #5 00b02410 in ?? ()
[22:18:55] [debug] #6 005d5cde in ?? () from plugins/streamer.so
[22:18:55] [debug] #7 005d66ac in ?? () from plugins/streamer.so
[22:18:55] [debug] #8 0060d9f0 in ?? () from plugins/streamer.so
[22:18:55] [debug] #9 0060e1aa in ?? () from plugins/streamer.so
[22:18:55] [debug] #10 005e0525 in ProcessTick () from plugins/streamer.so
[22:18:55] [debug] #11 080d1ce2 in ?? () from samp03svr
[22:18:55] [debug] #12 080aef6c in ?? () from samp03svr
[22:18:55] [debug] #13 080aa13a in ?? () from samp03svr
[22:18:55] [debug] #14 00259d26 in __libc_start_main () from /lib/libc.so.6
[22:18:55] [debug] #15 0804b4e1 in ?? () from samp03svr
The last two ocelots of each memory address ALWAYS remain the same when our server crashes, we are unsure if its an issue with streamer, what do you think?
We've removed crashdetect to see if the issue would go away, to no avail, the issue still existed.
The text was updated successfully, but these errors were encountered: