aMule Forum
English => aMule Help => Topic started by: Lucho[FLCL] on December 29, 2004, 08:51:04 PM
-
Hello everyone.
Last week I suddenly started getting lots of corrupted parts in aMule. I have to download ~ 3GB to get a complete 700 MB download.
I've tried recompiling aMule, recompiling amule along with it's dependencies (specially compression and hash checking stuff), changing the temp folder from one disk to another, running fsck in all the partitions, running memtest 86+ to see if it is a RAM problem, etc. Nothing worked.
So I end here looking for some ideas on what to try. aMule 2 is just around the corner, but I'm not sure if it will solve my problems.
Thanks in advance to all those that answer this post.
-
hi,
aMule 1.2.8 is a bit old ;) please update to 2.0 rc8
cheers
-
I'm not against moving to 2.0 rcX, but 1.2.8 is supposedly the last "release" quality aMule version, and it was working fine until last week.
I'll probably try aMule 2 rc8 if nothing else works....but I doubt if it will solve my problem...
I forgot to mention that I can download perfectly with BitTorrent, plain http, etc.
-
2.0 rc8 is a stable version for linux, like 1.2.8 is - so you don't have nothing to worry (rc is for mac osx) ;)
it is really recommended to update, because of partfile handling i.e .. and a lot more.
but sometimes, a file is really corrupted (wrong packed or smth like that) and u'll get those errors - there is nothing we can do about
cheers
-
Well, I'll install it tonight and post my results.
I know that sometimes a file is corrupted and nobody can do anything about it, but I'm getting that kind of error in every file I download; it completes and then it finds like 15 or so corrupted parts (repeat this behaviour 15x or so for most of the big files),.
I just posted because I didn't have this problem one week ago, and never had it with any other aMule release.
-
any problems with your filesystem / harddisk? maybe check dmesg or syslog
stefanero
-
No stefanero, I tried with 2 different disks, one being FAT32 and the other reiserfs3, and the problem persists.
-
ok...
I had a currpeted disk once, and had the same problem thats why I asked...
any improvement with rc8 ?
-
Originally posted by stefanero
ok...
I had a currpeted disk once, and had the same problem thats why I asked...
any improvement with rc8 ?
No, the problem persists.
I really don't know what to do. It may be a networking issue, because I reinstalled almost everything....
I have an ADSL connection....I'll ask my ISP just in case. But nonetheless, do you know if I can try anything else?
One strange thing is that if I download for example the kernel sources (~30MB) vía FTP or HTTP and check them with the MD5 sum, it passes without problems...Is there any socket library in the middle that may be corrupted? Just a thought.
BitTorrent downloads just fine (perhaps it gets corrupted parts and I don't notice because of it's high speed), I download files ~170MB without problems
-
Hmm...weird...this same problem has been discussed in the emule forums, since this very same problem happens sometimes with emule on windows. Solution there seems to be disabling temporarily ICH...I'm not sure if that's a solution or not, but it seems to work for some people. Still the problem is the causes for this are unknown. Search the emule forums, and see if you can get some applicable solution to this problem: http://forum.emule-project.net
Regards.
-
I wouldn't like to disable ICH right in the moment I'm getting corrupted parts :(
-
Originally posted by Lucho[FLCL]
I wouldn't like to disable ICH right in the moment I'm getting corrupted parts :(
Haha, OK, I get your feeling, but as I said, this has worked for emule before. Did you search in their forum for some other solutions?
Regards.
-
Originally posted by skolnick
Originally posted by Lucho[FLCL]
I wouldn't like to disable ICH right in the moment I'm getting corrupted parts :(
Haha, OK, I get your feeling, but as I said, this has worked for emule before. Did you search in their forum for some other solutions?
Regards.
I searched only lightly; I'll give it some more time later.
However, I don't expec to find much info regarding an aMule problem in the eMule forums... :P
I know that they are the same.....almost. There are a number of signifficant differences, and those differences are some of the candidates for my problems (some shared libraries for example...). However, as I said, I'll search more later
Any other ideas are appreciated.
-
do you use ipfilter Lucho[FLCL]?
-
I haven't cofigured it; unless it comes enabled by default. Why do you ask this?
-
because ipfilter block clients that are known for sending corrupted data ;)
-
Perhaps I'll try that, but I've got the feeling that I'll end up banning everyone in the net :D
-
I checked the config, and Ip Filtering is "on".
Thanks for the help so far, but it's still receiving corrupted parts. The messages are like the following one:
FileResultHash |33340A8661B7C9E1D69667CE2A443B35| FileHash |557955DEDAD30DFDA965ACB5F216779E|
12/30/04 17:10:57: Found corrupted part (70) in 76 parts file
Who do you say it's to blame?
Regards,
Luciano
-
Either the clients or your filesystem or the cryptopp lib.
Try recompiling amule not using the embedded crypto.
-
Originally posted by Kry
Either the clients or your filesystem or the cryptopp lib.
Try recompiling amule not using the embedded crypto.
First of all, thanks, it's really nice to see the devels and admins answering my thread :)
Back to the point: I'll give you my opinion on your thoughts:
1- I don't think that the clients are causing the problem, because I'm downloading different things and it's very unlikely that I get this many corrupted parts for every file. That, plus 1.2.8 was working fine till some day in the past week.
2- I blamed my filesystem too at first (the temp directory was on a FAT32 partition), so I changed the temp directory to another HD, which is reiserfs formatted. I deleted all the partial downloads and started them again, along with some new downloads, but the problem is still there. dmesg an syslog are clean, they are not reporting any kind of filesystem failure.
3- The crypto lib: How do I recompile aMule without it? I'm on gentoo, and that kind of thing is made via USE flags, these are the ones for amule: -debug -gd -gtk2 +nls -remote -stats -unicode. I don't see crypto stuff around. Is it something I should change in the makefile? Why do you think that this lib is to blame?
I think that the problem lies in some shared library, because I recompiled and even changed version of the app, changed the fs, checked the RAM, etc., and nothing worked.
Well, I really look forward to solving this problem ;). Thanks so far.
Luciano
EDIT: I recompiled aMule with less aggresive CFLAGS:
#CFLAGS="-O3 -mcpu=pentium3 -march=pentium3 -mmmx -msse -pipe"
But I'm still getting corrupted parts ?(
EDIT2: I don't know if it helps, but here is the output to ldd amule | egrep crypt (just because of the libcrypto thing that you mentioned)
bash-2.05b# ldd /usr/bin/amule | egrep crypt
libcrypto.so.0.9.7 => /usr/lib/libcrypto.so.0.9.7 (0xb7910000)
libcrypto.so.0.9.6 => /usr/lib/libcrypto.so.0.9.6 (0xb781e000)
EDIT3: Hey, the whole output is not so long, here it is, in case it helps:
bash-2.05b# ldd /usr/bin/amule
libpthread.so.0 => /lib/libpthread.so.0 (0x43e6b000)
libwx_gtk-2.4.so => /usr/lib/libwx_gtk-2.4.so (0xb7b3b000)
libgtk-1.2.so.0 => /usr/lib/libgtk-1.2.so.0 (0x43266000)
libgdk-1.2.so.0 => /usr/lib/libgdk-1.2.so.0 (0x439c7000)
libgmodule-1.2.so.0 => /usr/lib/libgmodule-1.2.so.0 (0x4428b000)
libglib-1.2.so.0 => /usr/lib/libglib-1.2.so.0 (0x43f03000)
libdl.so.2 => /lib/libdl.so.2 (0x43d4c000)
libXi.so.6 => /usr/X11R6/lib/libXi.so.6 (0xb7b32000)
libXext.so.6 => /usr/X11R6/lib/libXext.so.6 (0xb7b22000)
libX11.so.6 => /usr/X11R6/lib/libX11.so.6 (0xb7a54000)
libz.so.1 => /lib/libz.so.1 (0xb7a3e000)
libcurl.so.2 => /usr/lib/libcurl.so.2 (0x4432f000)
libssl.so.0.9.7 => /usr/lib/libssl.so.0.9.7 (0xb7a0d000)
libcrypto.so.0.9.7 => /usr/lib/libcrypto.so.0.9.7 (0xb7910000)
libstdc++.so.5 => /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/libstdc++.so.5 (0x43fb4000)
libm.so.6 => /lib/libm.so.6 (0x43d51000)
libgcc_s.so.1 => /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/libgcc_s.so.1 (0x4406f000)
libc.so.6 => /lib/libc.so.6 (0x43c1d000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x43c07000)
libgthread-1.2.so.0 => /usr/lib/libgthread-1.2.so.0 (0x43f73000)
libpng.so.3 => /usr/lib/libpng.so.3 (0x442d8000)
libjpeg.so.62 => /usr/lib/libjpeg.so.62 (0x43f87000)
libtiff.so.3 => /usr/lib/libtiff.so.3 (0x440e9000)
libssl.so.0.9.6 => /usr/lib/libssl.so.0.9.6 (0xb78df000)
libcrypto.so.0.9.6 => /usr/lib/libcrypto.so.0.9.6 (0xb781e000)
-
hi,
your ipfilter might be enabled, but do you have an ipfilter.dat file in your .aMule folder ?
-
Originally posted by deltaHF
hi,
your ipfilter might be enabled, but do you have an ipfilter.dat file in your .aMule folder ?
Yes, the file is there; however, it's empty ?(
-
http://www2.openmedia.info:8080/p27.html
unpack it, move then the ipfilter.dat file into your .aMule folder
happy new year & good night ;)
-
Originally posted by deltaHF
http://www2.openmedia.info:8080/p27.html
unpack it, move then the ipfilter.dat file into your .aMule folder
happy new year & good night ;)
Yesterday I recompiled openssl, curl, wxGTK, etc., with no success.
Has this anti-p2p war gone THIS far? Is it possible that my problem resides on fake clients?
Happy new year to you too. I'm trying your solution and will report my results later.
Thank you very much!
EDIT: Should I see any message in the log when a client is filtered by ipfilter?
-
Originally posted by Lucho[FLCL]
Originally posted by deltaHF
http://www2.openmedia.info:8080/p27.html
unpack it, move then the ipfilter.dat file into your .aMule folder
happy new year & good night ;)
Yesterday I recompiled openssl, curl, wxGTK, etc., with no success.
Has this anti-p2p war gone THIS far? Is it possible that my problem resides on fake clients?
Happy new year to you too. I'm trying your solution and will report my results later.
Thank you very much!
EDIT: Should I see any message in the log when a client is filtered by ipfilter?
I very sadly announce that I keep getting corrupted parts :(
The file completed 5 minutes ago, but 8 parts failed the hash check: 15-61-65-66-68-69-71-72
#65 was re-downloaded and failed the check again
I'll change kernel and see what happens....do you have any issues reported with 4KB stacks?
Edit: changed kernel (no 4kb stacks now), the new corrupted parts list is:
28-47-66-68-69-71 ?(
-
Hello everyone again. I hope you are enjoying the new year as much as I'm doing.
My problem is still there....I'm losing my hope to solve it...
I downloaded via plain HTTP several files just to see if I had some issue with my router/HDD. I downloaded several kernel sources and other files and md5'd them all, and the sums were all correct.
I recompiled several of the shared libraries that I found aMule used (with ldd, as I previously said).
So it's either my ISP screwing aMule traffic (I've got it running at another port, not the default), or some problem with the app (the hashing function?).
I installed an eDonkey server in my machine and connected two clients from different sessions to it, but I couldn't succeed in making one client download from the other (perhaps they are banned because of having the same IP? Just a thought). I wanted to do this test in order to discard problems with the ISP, but oh, well...
I'll wait for some other idea (what do I have to do to reinstall the WHOLE app?? Is it possible that some file in the .aMule folder is causing the problem? I wouldn't mind erasing the whole thing and starting all over. I don't mind losing my credits THAT much).
Thanks everyone for the help so far...I promise I won't post again if nobody answers...perhaps my problem is so obscure that is impossible to solve ;(
-
Lucho[FLCL],
Did you try testing your computer's memory? Maybe you have a hardware problem. Run memtest and leave it running for a while to see what happens.
Cheers!
-
Hmmm...I _know_ this sounds a lot like heresy, but, have you tried a different client? maybe xmule or mldonkey? it could (?) be possible that simply your system will not accept amule just like some people do not accept penicillin?? 8o
Regards.
Skolnick.
-
Originally posted by phoenix
Lucho[FLCL],
Did you try testing your computer's memory? Maybe you have a hardware problem. Run memtest and leave it running for a while to see what happens.
Cheers!
Yes, I did. It was the second or third thing I tried, memtest86+.
-
Originally posted by skolnick
Hmmm...I _know_ this sounds a lot like heresy, but, have you tried a different client? maybe xmule or mldonkey? it could (?) be possible that simply your system will not accept amule just like some people do not accept penicillin?? 8o
Regards.
Skolnick.
Hehehe perhaps you are right....I'm just sad because it was working fine and I really loved it.
However, I'm now testing an aMule compiled by hand (not by gentoo's portage) just to see what happens.
I appear to have wxGTK compiled against GTK2 and this is known to cause problems...could this be the cause? In portage I used USE="-gtk2" for aMule and it compiled fine, without warnings....but I don't know... ?(
-
well gtk2 is just know to have problems with high CPU usage and such, but not with screwing up the hashing....
which wxGTK version or you on? 2.4.2 or 2.5... something?
because if you are using 2.4.2 with gtk2 this would be really bad bad bad, high CPU and such...
and maybe the high CPU usage messes up your hashing ;)
stefanero
-
Originally posted by stefanero
well gtk2 is just know to have problems with high CPU usage and such, but not with screwing up the hashing....
which wxGTK version or you on? 2.4.2 or 2.5... something?
because if you are using 2.4.2 with gtk2 this would be really bad bad bad, high CPU and such...
and maybe the high CPU usage messes up your hashing ;)
stefanero
I monitor my CPU usage and it's never too high; I have a PIII 550 Mhz, so I would really notice it if it was happening.
Last night I noticed that the behaviour is really, really weird: when the message "downloaded part number xx is corrupt", the part isn't "deleted" from the progress bar, and is detected corrupted when the file is "completed" (along with others that weren't reported!!!). Is this the normal behaviour of the app?
The parts that are detected corrupt are always the same, plus others that vary from hashing to hashing.
The file is a movie, I tried to watch it using mplayer and I can see it OK. :O
-
pm me the link of that file plz
-
Originally posted by deltaHF
pm me the link of that file plz
OK, I will when I go home (I'm at work now), but in the meanwhile, just for you to know, I'm having the same problem with all the files ;)
-
Well, yesterday I recompiled wxGTK and aMule making sure everything is linked against GTK1 and not 2, but that didn't solve the problem. Next step: I'll try with another GNU/Linux installed somewhere else.
-
Originally posted by Lucho[FLCL]
Well, yesterday I recompiled wxGTK and aMule making sure everything is linked against GTK1 and not 2, but that didn't solve the problem. Next step: I'll try with another GNU/Linux installed somewhere else.
Well, I redownloaded the files I was downloading and I'm getting less corrupted parts (like 3 with the file being 80% complete). I'll repost when the files reach completion, but there seems to be light at the end of the tunnel ;)
-
*crosses fingers* ;)
-
Not sure if it still matters, but you should probably also try to recompile zlib with less aggressive flags. In fact, you would do yourself a favor if you did some reading on the problems with using aggressive optimization C(XX)Flags (at the Gentoo forum) and recompiled your entire system with something more constrained.