Scribe "kernel: potentially unexpected fatal signal 6" almost every second? (>50K entries)

[XIII]

Normally I don't inspect my router's system logfile that much, but lately the RAM drive on my Raspberry was running out of space, every couple of days. Each time the forwarded router logs were filling up that RAM drive. I still did not pay too much notice, but now that I finally do, I see that there are over 50K instances of the same error in the router log file. This error is logged almost every second:

Apr 24 14:10:48 kernel: potentially unexpected fatal signal 6.
Apr 24 14:10:48 kernel: CPU: 0 PID: 18973 Comm: syslog-ng Tainted: P           O    4.1.27 #2
Apr 24 14:10:48 kernel: Hardware name: Broadcom-v8A (DT)

What does this error mean?

 

[XIII]

Updating the Scribe filters seems to have helped masking the problem?

 

[SomeWhereOverTheRainBow]

Updating the Scribe filters seems to have helped masking the problem?

Whatever the case may be it has syslog-ng as apart of the log itself which begs to question is something going on with your syslogng setup? Specifically when anything Broadcom related restarts.

 

[XIII]

Looks like you're right.

After updating the Scribe filters the crash is not masked, as I thought; it no longer seems to occur!

Guess my syslog-ng configuration got borked somehow... (and fixed by updating the Scribe filters)

 

[SomeWhereOverTheRainBow]

Looks like you're right.

After updating the Scribe filters the crash is not masked, as I thought; it no longer seems to occur!

Guess my syslog-ng configuration got borked somehow... (and fixed by updating the Scribe filters)

Probably with the recent entware update. Alot of users noticed they has to update scribe immediately after.

 

[elorimer]

Not satisfied here.

I looked at the syslog-ng changes and didn't see anything that would affect us. Updating scribe would just change the version number in the config, but wouldn't change the filters. They haven't changed.

So what might be going on?

 

[SomeWhereOverTheRainBow]

Not satisfied here.

I looked at the syslog-ng changes and didn't see anything that would affect us. Updating scribe would just change the version number in the config, but wouldn't change the filters. They haven't changed.

So what might be going on?

well scribe does use some customization inside its launch files for taking control of the default syslog. without this initial modifications being made by scribe, syslog-ng would simply fail to start I imagine running the upgrade option from within scribe fixed those auxillary files necessary for starting syslog-ng too.

notice these two lines specifically

Apr 24 14:10:48 kernel: potentially unexpected fatal signal 6.
Apr 24 14:10:48 kernel: CPU: 0 PID: 18973 Comm: syslog-ng Tainted: P           O    4.1.27 #2

signal 6 is usually the signal to abort operations. From what it looks like, syslog-ng was failing to take over (a.k.a properly start). sometimes the config and filters files are bad, and sometimes the log files go corrupt while completing rotation. it is really hard to say what the issue might be.

Recently Entware just released updates on all the packages including syslog-ng. (this would modify the init.d launch file for syslog-ng if the user uses the terminal to upgrade instead of using scribe). The only way to fix the init.d file is to upgrade (or reinstall) using scribe menu. Scribe installs a customized init.d file for starting syslog-ng "properly".

 

[Rob Q]

I'm getting the same thing with Diverision Standard + pixelserv-tls.

Update: I think I found the source which might have been pixelserv in my case. I haven't played around with Scribe yet, maybe they both use the same process which causes the kernel errors.

Apr 27 09:55:58 kernel: dcd/9125: potentially unexpected fatal signal 11.
Apr 27 09:55:58 kernel: Pid: 9125, comm:                  dcd
Apr 27 09:55:58 kernel: CPU: 0    Tainted: P             (2.6.36.4brcmarm #1)
Apr 27 09:55:58 kernel: PC is at 0x401d4444
Apr 27 09:55:58 kernel: LR is at 0x1c070
Apr 27 09:55:58 kernel: pc : [<401d4444>]    lr : [<0001c070>]    psr: 20000010
Apr 27 09:55:58 kernel: sp : bee25a80  ip : 000a1b00  fp : bee25be4
Apr 27 09:55:58 kernel: r10: 00085f8c  r9 : 0000dafc  r8 : 00000068
Apr 27 09:55:58 kernel: r7 : 00085f8c  r6 : 00000000  r5 : bee25d34  r4 : 00000000
Apr 27 09:55:58 kernel: r3 : 00000000  r2 : 00000000  r1 : 00085f8c  r0 : 00000000
Apr 27 09:55:58 kernel: Flags: nzCv  IRQs on  FIQs on  Mode USER_32  ISA ARM  Segment user
Apr 27 09:55:58 kernel: Control: 10c53c7d  Table: 9b1c804a  DAC: 00000015

 

[SomeWhereOverTheRainBow]

I'm getting the same thing with Diverision Standard + pixelserv-tls.

Update: I think I found the source which might have been pixelserv in my case. I haven't played around with Scribe yet, maybe they both use the same process which causes the kernel errors.

Apr 27 09:55:58 kernel: dcd/9125: potentially unexpected fatal signal 11.
Apr 27 09:55:58 kernel: Pid: 9125, comm:                  dcd
Apr 27 09:55:58 kernel: CPU: 0    Tainted: P             (2.6.36.4brcmarm #1)
Apr 27 09:55:58 kernel: PC is at 0x401d4444
Apr 27 09:55:58 kernel: LR is at 0x1c070
Apr 27 09:55:58 kernel: pc : [<401d4444>]    lr : [<0001c070>]    psr: 20000010
Apr 27 09:55:58 kernel: sp : bee25a80  ip : 000a1b00  fp : bee25be4
Apr 27 09:55:58 kernel: r10: 00085f8c  r9 : 0000dafc  r8 : 00000068
Apr 27 09:55:58 kernel: r7 : 00085f8c  r6 : 00000000  r5 : bee25d34  r4 : 00000000
Apr 27 09:55:58 kernel: r3 : 00000000  r2 : 00000000  r1 : 00085f8c  r0 : 00000000
Apr 27 09:55:58 kernel: Flags: nzCv  IRQs on  FIQs on  Mode USER_32  ISA ARM  Segment user
Apr 27 09:55:58 kernel: Control: 10c53c7d  Table: 9b1c804a  DAC: 00000015

yes, your taint is different and involved dcd. the above users taint involves a bad syslog-ng run.

Your issue is a known issue and involves the use of the routers built in AiProtect in combination with trying to run pixelserv-tls. This issue is a closed source issue since the code comes from AiProtect side (a.k.a TrendMicro). Until Asus developers fix the issue inside the AiProtect, you will still see this present.