[opensuse-kernel] netconsole problems
I am looking for assistance in running netconsole. I am trying to run the statically linked version since I am trying to record the screen dump that occurs with crashes during boot. In summary, my netconsole appears to work fine, but it hangs before the output matches that presented by dmesg. If I run the module version, it behaves as expected printing messages from "modprobe floppy" and others suggested on the net. The static version is silent, until I issue a command to generate a panic, "echo c > /proc/sysrq-trigger". I have looked at the code and get lost trying to determine what is not happening. This is too early in the boot process and there are too many spinlocks. I assume that this is interrupt driven, but cannot isolate the logic. I am currently running opensuse 42.3 with a 4.18.7 kernel. This situation is repeatable. The command on the linux boot line includes: netconsole=64001@192.168.7.60/eth0,64001@192.168.7.55/34:e6:d7:01:2a:dd The command at my receiver is: sudo nc -v -u -l 64001 | tee netconsole.log AS indicated before, all the above appears to work until the hang, so I think my setup is correct. Is there anybody that would venture to experiment and see if this also happens to somebody else? I have tried asking on kernelnewbies and even sent an email to the maintainer, but have received no adequate response. Any suggestion would be appreciated. I will offer any other info I may have omitted. Don -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 10/10/2018, 01:13 AM, don fisher wrote:
Any suggestion would be appreciated.
Does adding "loglevel=7" to cmdline help? Anyway, upload dmesg somewhere. Netconsole would report about itself into the log. regards, -- js suse labs -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 10/9/18 10:43 PM, Jiri Slaby wrote:
On 10/10/2018, 01:13 AM, don fisher wrote:
Any suggestion would be appreciated.
Does adding "loglevel=7" to cmdline help?
Anyway, upload dmesg somewhere. Netconsole would report about itself into the log.
regards,
Below are the netconsole references contained in dmesg, followed by the lines that appear in dmesg that are not output by netconsole. Could the No iBFT detected message be fatal in some way and be killing netconsole? The command wc -c -m -l -L netconsole.log yields "769 55023 55023 294" stating the 769 lines were output with a total of 55,023 characters. Longest line is 294 characters. 2.315552 netpoll: netconsole: local port 64001 2.342735 netpoll: netconsole: local IPv4 address 192.168.7.60 2.418598 netpoll: netconsole: interface 'eth0' 2.443051 netpoll: netconsole: remote port 64001 2.467350 netpoll: netconsole: remote IPv4 address 192.168.7.55 2.491831 netpoll: netconsole: remote ethernet address 34:e6:d7:01:2a:dd 2.516548netpoll: netconsole: device eth0 not up yet, forcing it 5.505279 console [netcon0] enabled 5.532617] netconsole: network logging started The last line output by netconsole is: 12.384010] raid6: sse2x2 gen() 13237 MB/s The lines following the above in dmesg are: No iBFT detected. ieee80211 phy0: Selected rate control algorithm 'iwl-mvm-rs' thermal thermal_zone6: failed to read out thermal zone (-61) iwlwifi 0000:3d:00.0 wlan1: renamed from wlan0 raid6: sse2x4 gen() 15626 MB/s raid6: sse2x4 xor() 12119 MB/s raid6: avx2x1 gen() 26407 MB/s raid6: avx2x1 xor() 19139 MB/s raid6: avx2x2 gen() 29951 MB/s raid6: avx2x2 xor() 20916 MB/s raid6: avx2x4 gen() 33383 MB/s raid6: avx2x4 xor() 23620 MB/s raid6: using algorithm avx2x4 gen() 33383 MB/s raid6: .... xor() 23620 MB/s, rmw enabled raid6: using avx2x2 recovery algorithm xor: automatically using best checksumming function avx intel_rapl: Found RAPL domain package intel_rapl: Found RAPL domain core intel_rapl: Found RAPL domain uncore intel_rapl: Found RAPL domain dram Btrfs loaded, crc32c=crc32c-intel, assert=on Bluetooth: BNEP (Ethernet Emulation) ver 1.3 Bluetooth: BNEP filters: protocol multicast Bluetooth: BNEP socket layer initialized BTRFS: device label os62 devid 1 transid 11544 /dev/nvme1n1p2 NET: Registered protocol family 17 NFSD: starting 90-second grace period (net f00000a8) I can sent the dumps from netconsole and dmesg if desired. Includes make for log email messages. Don -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Wed 2018-10-10 14:36:29, don fisher wrote:
On 10/9/18 10:43 PM, Jiri Slaby wrote:
On 10/10/2018, 01:13 AM, don fisher wrote:
Any suggestion would be appreciated.
Does adding "loglevel=7" to cmdline help?
Anyway, upload dmesg somewhere. Netconsole would report about itself into the log.
regards,
Below are the netconsole references contained in dmesg, followed by the lines that appear in dmesg that are not output by netconsole. Could the No iBFT detected message be fatal in some way and be killing netconsole? The command wc -c -m -l -L netconsole.log yields "769 55023 55023 294" stating the 769 lines were output with a total of 55,023 characters. Longest line is 294 characters.
2.315552 netpoll: netconsole: local port 64001 2.342735 netpoll: netconsole: local IPv4 address 192.168.7.60 2.418598 netpoll: netconsole: interface 'eth0' 2.443051 netpoll: netconsole: remote port 64001 2.467350 netpoll: netconsole: remote IPv4 address 192.168.7.55 2.491831 netpoll: netconsole: remote ethernet address 34:e6:d7:01:2a:dd 2.516548netpoll: netconsole: device eth0 not up yet, forcing it 5.505279 console [netcon0] enabled 5.532617] netconsole: network logging started
The last line output by netconsole is: 12.384010] raid6: sse2x2 gen() 13237 MB/s
The lines following the above in dmesg are: No iBFT detected. ieee80211 phy0: Selected rate control algorithm 'iwl-mvm-rs' thermal thermal_zone6: failed to read out thermal zone (-61) iwlwifi 0000:3d:00.0 wlan1: renamed from wlan0
I wonder if this might somehow affect the networking interfaces. I am not sure what the above forcing of eth0 interface up really means and if it might later be modified. Note that I neither have much experiences with netconsole not I am networking guy. I am just interested into reasons where any console stop working.
I can sent the dumps from netconsole and dmesg if desired. Includes make for log email messages
Full logs are always better. You could eventually share them via http://paste.suse.de/ or bugzilla. Best Regards, Petr -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Thursday, 11 October 2018 9:36 Petr Mladek wrote:
Full logs are always better. You could eventually share them via http://paste.suse.de/ or bugzilla.
Rather paste.opensuse.org, paste.suse.de is not public. Possibly related: http://lkml.kernel.org/r/e0611057-1b5b-9990-b9fa-060a9b5eba40@linux.ee but it's hard to say without more information. Michal Kubecek -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 10/11/18 12:47 AM, Michal Kubecek wrote:
On Thursday, 11 October 2018 9:36 Petr Mladek wrote:
Full logs are always better. You could eventually share them via http://paste.suse.de/ or bugzilla.
Rather paste.opensuse.org, paste.suse.de is not public.
Possibly related:
http://lkml.kernel.org/r/e0611057-1b5b-9990-b9fa-060a9b5eba40@linux.ee
but it's hard to say without more information.
Michal Kubecek
Thanks for the reference. I cannot tell from this output if it is a similar problem. How does one manage to trigger the "Call Trace" shown at time 9.914818? Again, mine works through a lot of the boot, then hangs. The "netconsole: network logging started" in this dump is listed at 8.788898, while mine occurs at 5.532617. Probably not significant. Please advise what to place on paste.opensuse.org, just dmesg, output from netconsole or both. Don -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 10/11/18 12:47 AM, Michal Kubecek wrote:
On Thursday, 11 October 2018 9:36 Petr Mladek wrote:
Full logs are always better. You could eventually share them via http://paste.suse.de/ or bugzilla.
Rather paste.opensuse.org, paste.suse.de is not public.
Possibly related:
http://lkml.kernel.org/r/e0611057-1b5b-9990-b9fa-060a9b5eba40@linux.ee
but it's hard to say without more information.
Michal Kubecek
How does one get a copy of the text displayed on the boot screen into dmesg dump or some other file? My verbosity flags are set at 8. There are messages like "Reached target basic system" and "Started System Logging Service" that are on my console screen that do not appear in the dmesg dump. I wrongly assumed that what appeared in the dmesg dump was what was shown on the console. Don -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 10/11/18 12:47 AM, Michal Kubecek wrote:
On Thursday, 11 October 2018 9:36 Petr Mladek wrote:
Full logs are always better. You could eventually share them via http://paste.suse.de/ or bugzilla.
Rather paste.opensuse.org, paste.suse.de is not public.
Possibly related:
http://lkml.kernel.org/r/e0611057-1b5b-9990-b9fa-060a9b5eba40@linux.ee
but it's hard to say without more information.
Michal Kubecek
Noticed from /var/log/messages, that: systemd[1]: Started wicked AutoIPv4 supplicant service. appears right after output terminates. Could wicked be killing netconsole? Don -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
participants (4)
-
don fisher -
Jiri Slaby -
Michal Kubecek -
Petr Mladek