gday all
have found that the daemon is stopping within a few minutes if there is more than 1 parralel download active. runs fine if set to just 1 download.
bug is repeatable and logfile shows nothing out of the ordinary, daemon just stops existing & has to be restarted?
im using the latest lucid nightly build thru apt. plenty of memory, drive space etc.
(i no longer have the facility to compile from svn to see if it is in the latest code)
all downloads are rapid premium
let me know if u need any more info
cheers :)
Hey there,
1: You are sure you don't get a "CRASH DETECTED" in your logs?
2: On which platform? Linux/Windows? ARM/MIPS/x86?
3: What kind of links? plugin-related? Or just plain http links?
Adrian
gday adrian
sorry for the delay in replying.
i get nothing in the log at all....
Mar 23 20:31:55 server DownloadDaemon[1195]: DEBUG: Successfully parsed download ID: 0
Mar 23 20:31:55 server DownloadDaemon[1195]: DEBUG: Starting download ID: 0
Mar 23 20:31:56 server DownloadDaemon[1195]: DEBUG: Successfully parsed download ID: 1
Mar 23 20:31:56 server DownloadDaemon[1195]: DEBUG: Starting download ID: 1
Mar 23 20:31:57 server DownloadDaemon[1195]: DEBUG: Successfully parsed download ID: 2
Mar 23 20:31:57 server DownloadDaemon[1195]: DEBUG: Starting download ID: 2
Mar 23 20:31:58 server DownloadDaemon[1195]: DEBUG: Successfully parsed download ID: 3
Mar 23 20:31:58 server DownloadDaemon[1195]: DEBUG: Starting download ID: 3
and thats the last entries in the log. last write to the ".part" files was 3 minutes later and deamon no longer existed after that.
i did some testing on the weekend and found that downloading from my hotfile premium account did not have this problem, only happens when downloading from my rapidshare premium??? downloading just 1 file at a time from rs prem works perfect still.
system is x86 running linux ubuntu lucid server 32bit, 4G ram, headless
cheers :)
steve
I'm using build 1051 on my WD Live.
I have a RS premium account and it's running fine with 3 parallel downloads. So I would say no general RS plugin error.
gday all
bug still exists in nightly build 1151 unfortunately. if theres anything more i can do to help track it down, let me know.
cheers :)
I still can't think of any reason why this could happen.
Never seen this bug and you are the only one who seems to be affected by it.
If you don't mind the hassle, try this:
1: Check out DD from SVN and compile/install it. When compiling, pass this flag to cmake: "-DCMAKE_CXX_FLAGS=-g"
2: stop all DD instances and start DD in a debugger like that:
gdb /usr/local/bin/DownloadDaemon
3: Type "run" in the gdb-shell
4: try to get DD to crash
5: after DD crashed, gdb should tell you where it happened. type "bt" in the gdb-shell to get the exact location/callstack
6: post that here..
Adrian
Or set the debug to: file: /var/downloads/downloaddaemon.log => will also show why downloaddaemon has crashed!
nope. It usually should show that with the CRASH DETECTED line. But that should also work with the other logging options.
What I need is a debugging-backtrace, since OP obviously doesn't get the CRASH DETECTED line
Adrian
havent got the time to compile till the weekend but i tried the debugging thing with the installed program... this is what i got when it crashed........
DownloadDaemon: ath.c:193: _gcry_ath_mutex_lock: Assertion `*lock == ((ath_mutex_t) 0)' failed.
Program received signal SIGABRT, Aborted.
[Switching to Thread 0xb56feb70 (LWP 14711)]
0xb7fe2430 in __kernel_vsyscall ()
(gdb) bt
0 0xb7fe2430 in __kernel_vsyscall ()
1 0xb7d1b651 in raise () from /lib/tls/i686/cmov/libc.so.6
2 0xb7d1ea82 in abort () from /lib/tls/i686/cmov/libc.so.6
3 0xb7d14718 in __assert_fail () from /lib/tls/i686/cmov/libc.so.6
4 0xb7b04675 in ?? () from /lib/libgcrypt.so.11
5 0xb7b40e60 in ?? () from /lib/libgcrypt.so.11
6 0xb7b41b6c in ?? () from /lib/libgcrypt.so.11
7 0xb7b40a80 in ?? () from /lib/libgcrypt.so.11
8 0xb7b10d37 in ?? () from /lib/libgcrypt.so.11
9 0xb7b10ee7 in ?? () from /lib/libgcrypt.so.11
10 0xb7b1142d in ?? () from /lib/libgcrypt.so.11
11 0xb7b11492 in ?? () from /lib/libgcrypt.so.11
12 0xb7afb324 in gcry_md_read () from /lib/libgcrypt.so.11
13 0xb7ba2e8d in ?? () from /usr/lib/libgnutls.so.26
14 0xb7b8508b in ?? () from /usr/lib/libgnutls.so.26
15 0xb7b850b9 in ?? () from /usr/lib/libgnutls.so.26
16 0xb7b784a1 in ?? () from /usr/lib/libgnutls.so.26
17 0xb7b7897e in ?? () from /usr/lib/libgnutls.so.26
18 0xb7b76250 in _gnutls_recv_int () from /usr/lib/libgnutls.so.26
19 0xb7b76cfb in gnutls_record_recv () from /usr/lib/libgnutls.so.26
20 0xb7fbb806 in ?? () from /usr/lib/libcurl-gnutls.so.4
21 0xb7fbcaad in ?? () from /usr/lib/libcurl-gnutls.so.4
22 0xb7f9a74f in ?? () from /usr/lib/libcurl-gnutls.so.4
23 0xb7fad9f9 in ?? () from /usr/lib/libcurl-gnutls.so.4
24 0xb7fafa15 in ?? () from /usr/lib/libcurl-gnutls.so.4
25 0xb7fb05a3 in curl_easy_perform () from /usr/lib/libcurl-gnutls.so.4
26 0x080689d0 in ddcurl::perform() ()
27 0x0808673f in download::download_me_worker(dl_cb_info&) ()
28 0x080881bc in download::download_me() ()
29 0xb7f45c96 in ?? () from /usr/lib/libstdc++.so.6
30 0xb7cd396e in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
31 0xb7dbea4e in clone () from /lib/tls/i686/cmov/libc.so.6
Uhm.. to me, this looks more like an error in your CURL library.
Did you install it manually or is it the one that ships with ubuntu?
Maybe you need to just update curl..
everything is standard ubuntu server install or thru apt. (no facility to compile other software yet)
curl is 7.19.7-1ubuntu1.1
cheers
Googling for "ath.c:193: _gcry_ath_mutex_lock: Assertion `*lock ==
((ath_mutex_t) 0)' failed" turns up a whole bunch of results. It seems to be a problem in the gnutls library shipped with ubuntu lucid.
My recommendation would be compiling libcurl manually using openssl instead of gnutls and then compiling/linking DownloadDaemon against that libcurl version..
Or just don't download from multiple ssl-encrypted sites in parallel :/
Note: You can try again with rev. 1177. But I doubt that it will change anything.
Adrian