Hi FreeIPA users,
I've been haunted across installs by a sporadic krb5kdc segfault, the especially fun part is that it seems to bring the service down on all of the servers at once! Restarting it brings everything back again quite happily..
The last and only useful krb5kdc.log entry is:
Nov 29 12:12:13 ipa1.redacted.net krb5kdc[20700](Error): worker 20705 exited with status 139 and in terms of segfault related, dmesg: [271761.231312] krb5kdc[20705]: segfault at fffffffffffffff8 ip 00007fe492d56b80 sp 00007ffc55c38c80 error 5 in ipadb.so[7fe492d4c000+19000]
I was hoping that moving from our Fedora based install to CentOS would make this go away, but it seems to have followed us!
Any advice as to where best to seek more useful log data?
Best,
David
On pe, 29 marras 2019, David Harvey via FreeIPA-users wrote:
Hi FreeIPA users,
I've been haunted across installs by a sporadic krb5kdc segfault, the especially fun part is that it seems to bring the service down on all of the servers at once! Restarting it brings everything back again quite happily..
The last and only useful krb5kdc.log entry is:
Nov 29 12:12:13 ipa1.redacted.net krb5kdc[20700](Error): worker 20705 exited with status 139 and in terms of segfault related, dmesg: [271761.231312] krb5kdc[20705]: segfault at fffffffffffffff8 ip 00007fe492d56b80 sp 00007ffc55c38c80 error 5 in ipadb.so[7fe492d4c000+19000]
I was hoping that moving from our Fedora based install to CentOS would make this go away, but it seems to have followed us!
Any advice as to where best to seek more useful log data?
If you would install debuginfo packages for ipa / krb5, and enable abrt processing, we might get a whole backtrace.
There were some fixes in ipadb.so recently in upstream.
Thanks for the swift response Alexander. I'll try and get that enabled for clearer details.
On Fri, 29 Nov 2019 at 13:59, Alexander Bokovoy abokovoy@redhat.com wrote:
On pe, 29 marras 2019, David Harvey via FreeIPA-users wrote:
Hi FreeIPA users,
I've been haunted across installs by a sporadic krb5kdc segfault, the especially fun part is that it seems to bring the service down on all of the servers at once! Restarting it brings everything back again quite happily..
The last and only useful krb5kdc.log entry is:
Nov 29 12:12:13 ipa1.redacted.net krb5kdc[20700](Error): worker 20705 exited with status 139 and in terms of segfault related, dmesg: [271761.231312] krb5kdc[20705]: segfault at fffffffffffffff8 ip 00007fe492d56b80 sp 00007ffc55c38c80 error 5 in
ipadb.so[7fe492d4c000+19000]
I was hoping that moving from our Fedora based install to CentOS would
make
this go away, but it seems to have followed us!
Any advice as to where best to seek more useful log data?
If you would install debuginfo packages for ipa / krb5, and enable abrt processing, we might get a whole backtrace.
There were some fixes in ipadb.so recently in upstream.
-- / Alexander Bokovoy Sr. Principal Software Engineer Security / Identity Management Engineering Red Hat Limited, Finland
freeipa-users@lists.fedorahosted.org