Univention Bugzilla – Bug 25918
Unterstützung für crash dump kernel
Last modified: 2020-07-04 16:41:16 CEST
Für die Fehlersuche beim Kunden (insbesondere 26) wäre es hilfreich, wenn in unseren Kernel die Option CONFIG_CRASH_DUMP=y aktiviert wäre. Dies erlaubt es einen sog. "Crash-Dump-Kernel" in den Speicher zu laden, der im Falle eines Kernel-Crashes die Kontrolle über das System übernimmt und dann zur Fehlerdiagnose genutzt werden kann, z.B. in dem dieser den Zustand des gecrashten Systems auf die Platte sichert. <http://www.kernel.org/doc/Documentation/kdump/kdump.txt> CONFIG_KEXEC=y # bereits aktiviert CONFIG_SYSFS=y # bereits aktiviert CONFIG_DEBUG_INFO=Y # bereits aktiviert CONFIG_CRASH_DUMP=y # neu CONFIG_PROC_VMCORE=y # neu CONFIG_RELOCATABLE=y # bereits aktiviert
# egrep -H '^CONFIG_(KEXEC|SYSFS|DEBUG_INFO|CRASH_DUMP|PROC_VMCORE|RELOCATABLE)\>' /boot/config-`uname -r` /boot/config-4.1.0-ucs153-amd64:CONFIG_KEXEC=y /boot/config-4.1.0-ucs153-amd64:CONFIG_CRASH_DUMP=y /boot/config-4.1.0-ucs153-amd64:CONFIG_RELOCATABLE=y /boot/config-4.1.0-ucs153-amd64:CONFIG_PROC_VMCORE=y /boot/config-4.1.0-ucs153-amd64:CONFIG_SYSFS=y /boot/config-4.1.0-ucs153-amd64:CONFIG_DEBUG_INFO=y # egrep -H '^CONFIG_(KEXEC|SYSFS|DEBUG_INFO|CRASH_DUMP|PROC_VMCORE|RELOCATABLE)\>' /boot/config-`uname -r` /boot/config-3.16.0-ucs189-amd64:CONFIG_KEXEC=y /boot/config-3.16.0-ucs189-amd64:CONFIG_CRASH_DUMP=y /boot/config-3.16.0-ucs189-amd64:CONFIG_RELOCATABLE=y /boot/config-3.16.0-ucs189-amd64:CONFIG_PROC_VMCORE=y /boot/config-3.16.0-ucs189-amd64:CONFIG_SYSFS=y /boot/config-3.16.0-ucs189-amd64:CONFIG_DEBUG_INFO=y SDB-Article <http://sdb.univention.de/1362> Also see <https://wiki.ubuntu.com/Kernel/CrashdumpRecipe>
Created attachment 7535 [details] UCS kdump Setup crash dump kernel and collector: ./25918_kexec-crash "collector.$(dnsdomainname)"
Created attachment 8041 [details] UCS kdump v2 Setup crash dump kernel and collector: ./25918_kexec-crash "collector.$(dnsdomainname)"
Created attachment 8047 [details] UCS kdump v3 Don't load next kernel by default Fix `date +%%s`
Created attachment 8048 [details] UCS kdump v4 After kexec-reboot the VGA console is broken because the VESA-mode is not saved/restored correctly. You should disable the VESA-FB for Linux: ucr unset grub/gfxmode grub/gfxpayload
Created attachment 8055 [details] UCS kdump v5 Add options to specify --interface,--driver in case auto-detect fails (bridge,bonding,...) --size in case 128 MiB is not enough - lots of kernel modules? Tries all interfaces to reach crash collect server Output debug info on console. Fixes sysctl syntax.
This issue has been filed against UCS 3. UCS 3 is out of the normal maintenance and many UCS components have vastly changed in UCS 4. If this issue is still valid, please change the version to a newer UCS version otherwise this issue will be automatically closed in the next weeks.
There is a Customer ID set so I set the flag "Enterprise Customer affected".
Created attachment 9219 [details] UCS kdump v6 Fix $output in collector.sh not being expanded. Detect Xen hypervisor.
Created attachment 9289 [details] UCS kdump v7 Add bridge detection
Created attachment 9298 [details] UCS kdump v8 Fix IP detection Add --force option Ignore more kernel command line arguments
This issue has been filed against UCS 4.2. UCS 4.2 is out of maintenance and many UCS components have changed in later releases. Thus, this issue is now being closed. If this issue still occurs in newer UCS versions, please use "Clone this bug" or reopen it and update the UCS version. In this case please provide detailed information on how this issue is affecting you.
We switched to upstream Debian kernels, which have it enabled by default: $ grep CONFIG_CRASH_DUMP /boot/config-4.19.0-9-amd64 CONFIG_CRASH_DUMP=y