When trying to login using WebAccess with a user that doesn't exist on the server you get a delay of ~50 seconds before it says authentication failed. It seems to be selinux related since `setenforce 0` fixes the problem. The audit log says: ==> /var/log/audit/audit.log <== type=USER_AVC msg=audit(1530090502.684:1025): pid=776 uid=81 auid=4294967295 ses=4294967295 subj=system_u:system_r:system_dbusd_t:s0-s0:c0.c1023 msg='avc: denied { send_msg } for msgtype=error error_name=org.freedesktop.systemd1.NoSuchDynamicUser dest=:1.248 spid=1 tpid=14879 scontext=system_u:system_r:init_t:s0 tcontext=system_u:system_r:thinlinc_webaccess_t:s0 tclass=dbus permissive=0 exe="/usr/bin/dbus-daemon" sauid=81 hostname=? addr=? terminal=?' The delay problem does not exist when logging in to the same system using the same non-existing user with SSH. You get a different AVC in the audit log: ==> /var/log/audit/audit.log <== type=USER_AVC msg=audit(1530089623.012:952): pid=776 uid=81 auid=4294967295 ses=4294967295 subj=system_u:system_r:system_dbusd_t:s0-s0:c0.c1023 msg='avc: denied { send_msg } for msgtype=method_call interface=org.freedesktop.systemd1.Manager member=LookupDynamicUserByName dest=org.freedesktop.systemd1 spid=13483 tpid=1 scontext=system_u:system_r:sshd_t:s0-s0:c0.c1023 tcontext=system_u:system_r:init_t:s0 tclass=dbus permissive=0 exe="/usr/bin/dbus-daemon" sauid=81 hostname=? addr=? terminal=?' Problem was first encountered on Fedora 27, can reproduce on a different Fedora 28 system as well. Can not reproduce on eudemo or tl.cendio.se.
This might be related: https://bugzilla.redhat.com/show_bug.cgi?id=1460244
Before bug 5086: We always (on these systems) had a very long (~75 seconds) delay after trying to login with a user that didn't exist. Then we eventually get a "Authentication failure" error in the GUI. Logs a normal auth-fail, no hints to why it took so long, eg. no timeout reached due to 120 secs login grace time. After bug 5086: We properly handle errors in PAM conversation and get a timeout after 10 seconds and a "Internal error" in the GUI. Logs say that a timeout was reached in the PAM conversation. -- The difference is now that the admin gets a hint that something is wrong with his system.
Also see bug 7277
The root cause was found in bug 7277. The SELinux policy is a bit broken and doesn't allow things to send dbus replies back to our unconfined processes. And nss_systemd uses dbus to work, so whenever that module was used we get a hang until dbus times out. The reason it only happens on invalid users is because nss_systemd is the last module in most configurations, so valid users will be found by one of the earlier modules. This will be fixed as a side effect of bug 7277 being fixed.
Works well now, tested with build 6040 on Fedora 29. Release note look good as well.