4:04 a.m.
Thanks for your help!
This is coming from an attempt to get a Kerberos service ticket
using
credentials for the user you are using to enroll this machine. Since you
are passing '-w$password' and not any specific principal, this means it
is the machine itself, hence we see
I'm passing -phost-enrollment (host-enrollment is the user for the password in -w),
should I be adding something more?
'TGT has been revoked' error comes from your KDC on IPA
master. Please check
/var/log/krb5kdc.log on IPA server you connected to for this deployment.
There should be one of explaining messages prior to rejection. It might
be prefixed with 'PAC issue:' string
There's nothing around this exact time, the only bits regarding a specific failed host
that I could find are:
krb5kdc[4526](info): AS_REQ (6 etypes {aes256-cts-hmac-sha1-96(18),
aes256-cts-hmac-sha384-192(20), camellia256-cts-cmac(26), aes128-cts-hmac-sha1-96(17), a
es128-cts-hmac-sha256-128(19), camellia128-cts-cmac(25)}) 172.22.2.123: NEEDED_PREAUTH:
host/ip-172-22-2-123.xxx(a)EXAMPLE.COM for krbtgt/EXAMPLE.COM(a)EXAMPLE.COM, Additional
pre-authentication required
krb5kdc[4526](info): closing down fd 4
krb5kdc[4525](info): AS_REQ (6 etypes {aes256-cts-hmac-sha1-96(18),
aes256-cts-hmac-sha384-192(20), camellia256-cts-cmac(26), aes128-cts-hmac-sha1-96(17),
aes128-cts-hmac-sha256-128(19), camellia128-cts-cmac(25)}) 172.22.2.123: ISSUE: authtime
1669105826, etypes {rep=aes256-cts-hmac-sha1-96(18), tkt=aes256-cts-hmac-sha1-96(18),
ses=aes256-cts-hmac-sha1-96(18)}, host/ip-172-22-2-123.xxx(a)EXAMPLE.COM for
krbtgt/EXAMPLE.COM(a)EXAMPLE.COM
(Different host in the logs above than before, but we had this issue again this morning
and it was easier to look up that older logs).
In terms of errors, the only one I could find in the logs was:
krb5kdc[27486](Error): PAC issue: ipadb_get_principal failed.
This was logged right before the whole set of instances got their errors.
I did also notice that the some of the same hostnames exist in older Kerberos logs
(hostnames will get repeated in our cloud env every now and then), could this be the
cause? A host with a previously used hostname trying to enrol again? We have an automated
process in place that calls host-del to IPA when an instance is terminated to delete it
and its data from IPA, but maybe we should be clearing something from Kerberos directly
too?