Admin console not accessable after 15-1 update

[fvdw]

A job for this evening.
When it froze I still could access via SSH de NAS but not all commands work anymore.
I forgot to try the dmesg command

Next that it freezes on your nas, try to logging via ssh and try dmesg command. If it is a kernel oops or panic and dmesg list it then it will be very helpful to find the cause

[brinka123]

After 1day nzb crashed.
This is the dmesg output:
(I PM'd a more complete output)

Code: Select all

[92216.079793] Code: e08ff00c e1a00000 e1a00000 e4803004 (e4804004)
[92216.086212] ---[ end trace 19099c8b756ddcc8 ]---
[92216.090891] note: nzbget[8287] exited with preempt_count 1
[92216.096390] Unable to handle kernel paging request at virtual address 00084c3                                       0
[92216.103623] pgd = cf288000
[92216.106317] [00084c30] *pgd=0f394831, *pte=03ad314f, *ppte=03ad3ffe
[92216.112610] Internal error: Oops: 817 [#4] PREEMPT ARM
[92216.117727] Modules linked in: iscsi_tcp libiscsi_tcp libiscsi scsi_transport                                       _iscsi iscsi_trgt(O) usblp usb_storage ehci_hcd
[92216.128973] CPU: 0    Tainted: G      D    O  (3.9.5 #199)
[92216.134436] PC is at memcpy+0xc4/0x3a4
[92216.138173] LR is at 0x666f3339
[92216.141305] pc : [<c02d9044>]    lr : [<666f3339>]    psr: 20000013
[92216.141305] sp : cf199ce4  ip : 00000010  fp : cf198000
[92216.152731] r10: 00000400  r9 : cb7d3084  r8 : 34747261
[92216.157934] r7 : 703c2030  r6 : 20303232  r5 : 00084c30  r4 : c02d900c
[92216.164431] r3 : 00000010  r2 : 000003a0  r1 : cb7d3094  r0 : 00084c30
[92216.170929] Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
[92216.178031] Control: 0005317f  Table: 0f304000  DAC: 00000015
[92216.183749] Process nzbget (pid: 8284, stack limit = 0xcf1981b8)
[92216.189728] Stack: (0xcf199ce4 to 0xcf19a000)


[brinka123]

Look at this:

http://archlinuxarm.org/forum/viewtopic.php?f=53&t=3828

[fvdw]

mmm..not an easy one to analyze

did you do the test with kernel 164 ?

[brinka123]

Test result is with the latest 15-1 kernel....

[fvdw]

I did not get a crash yet with he 164 kernel further I will recompile nzbget with debug enabled maybe that gives more information.
I will post it in a minute
From the oops error output and links on the internet I did not find a solution yet

---edit
attached the nzbget version with debug enable, it will create a log file in the dst subfolder in the download folder you defined.

put this nzbget in fvdw share in the addons/bin folder replacing the existings nbzget binary be sure that it keeps executable permissions otherwise it won't run

[fvdw]

debug is not given away the secret. This were the last lines in nzbget.log when the nas crashed (hangs)

Code: Select all

on Mar 17 23:48:30 2014   2941252704   DEBUG   Destroying Decoder (Decoder.cpp:61:~Decoder)
Mon Mar 17 23:48:30 2014   2974807136   DEBUG   Entering Thread-func (Thread.cpp:248:thread_handler)
Mon Mar 17 23:48:30 2014   2974807136   DEBUG   Entering ArticleDownloader-loop (ArticleDownloader.cpp:124:Run)
Mon Mar 17 23:48:30 2014   2974807136   DEBUG   Opening connection to news.onsbrabantnet.nl (NNTPConnection.cpp:224:Connect)
Mon Mar 17 23:48:30 2014   2974807136   DETAIL   Downloading Skyfall.2012.DVDRiP.XviD-BjX/Skyfall.2012.DVDRiP.XviD-BjX.part44.rar [10/28] @ server1 (news.onsbrabantnet.nl)
Mon Mar 17 23:48:30 2014   2941252704   DEBUG   Destroying Thread (Thread.cpp:174:~Thread)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Releasing connection (ArticleDownloader.cpp:905:FreeConnection)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Freeing used connection (ServerPool.cpp:254:FreeConnection)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Expected crc32=e4408cbb (Decoder.cpp:332:Check)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Calculated crc32=e4408cbb (Decoder.cpp:333:Check)
Mon Mar 17 23:48:30 2014   3016750176   DETAIL   Successfully downloaded Skyfall.2012.DVDRiP.XviD-BjX/Skyfall.2012.DVDRiP.XviD-BjX.part44.rar [7/28]
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Notifying observers (Observer.cpp:55:Notify)
Mon Mar 17 23:48:30 2014   3054236768   DEBUG   Starting new ArticleDownloader (QueueCoordinator.cpp:624:StartArticleDownload)
Mon Mar 17 23:48:30 2014   3054236768   DEBUG   Creating Thread (Thread.cpp:159:Thread)
Mon Mar 17 23:48:30 2014   3054236768   DEBUG   Creating Decoder (Decoder.cpp:52:Decoder)
Mon Mar 17 23:48:30 2014   3054236768   DEBUG   Creating Decoder (Decoder.cpp:52:Decoder)
Mon Mar 17 23:48:30 2014   3054236768   DEBUG   Creating ArticleDownloader (ArticleDownloader.cpp:62:ArticleDownloader)
Mon Mar 17 23:48:30 2014   3054236768   DEBUG   Starting Thread (Thread.cpp:182:Start)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Notification from ArticleDownloader received (QueueCoordinator.cpp:658:Update)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Article downloaded (QueueCoordinator.cpp:671:ArticleCompleted)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Exiting ArticleDownloader-loop (ArticleDownloader.cpp:338:Run)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Thread-func exited (Thread.cpp:254:thread_handler)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Autodestroying Thread-object (Thread.cpp:260:thread_handler)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Destroying ArticleDownloader (ArticleDownloader.cpp:78:~ArticleDownloader)
Mon Mar 17 23:48:30 2014   3016750176   DEBUG   Destroying Decoder (Decoder.cpp:61:~Decoder)

pretty similar as yours. nzbget is doing something wrong in the system memory. But it seems to be random...difficult

to be continued. Tomorrow I set the nwsp2 with the 164 kernel, now its time to go to bed

the normal approach to insert in the suspected routine of nzbget some print statement to find out were the code fails

[brinka123]

Other approach is to check the code with a code checker.
It will reveal nasty code.
But still the kernel should be robust against this.
Did you read about the meaning of the Oops messages?
817 is an hexadecal. The bits have a meaning and explain the kernel error.

[fvdw]

probably the kernel code is ok and is the code of nzbget causing the problem by making some illegal operations or operations not compatible with latest kernel code.

I tried to find the the menaing of the oops error (in my case it mentioned 805 and not 817) but did not find a clear explanantion other than various reasons

still the question, did you try again kernel 164?
this evening I will run my nwsp with 164 again to see if it crashes with nzbget running and downloading
If you do not use download then nzbget doesn't cause a problem, so the bug is related to the download routines used by nzbget

[brinka123]

Running with kernel 164 now. Will report results...

I am also convinced it happens during downloading (not during parchecking and repairing).

Did you read this?
http://www.armadeus.com/wiki/index.php?title=Oops%2C_Linux_crashes
Can't install ksymoops. apt-get doesnt work.... It will exactly tell in which kernel function the error occurs.
Do you have a System.map (compilation result i think...) in this file we can find adresses of the kernel function calls. This adresses are reported in de dmesg message...


Regards and thanks for your investing time in this.