Dear Maintainer,


my computer receive DNS IPv6 RA updates and maintain resolv.conf using rdnssd.
Ra updates are frequent so the resolv.conf mtime is updated every few seconds.
sssd then load the computer up to 50% of cpu usage due to this change, even
there is no content change in resolv.conf.

Stopping rdnssd prevent sssd from producing this high cpu usage.

I'm not sure if the bug is that sssd should check for content change of
resolv.conf or if rdnssd shouldn't update it because RA received contains no
change. So feel free to reassign if needed.

I tried this with jessie sssd and with a backport of sid one, no change.

here is my ssd.conf for instance:


[sssd]
config_file_version = 2
reconnection_retries = 3
sbus_timeout = 30
services = nss, pam
domains = foo.bar

[nss]
filter_groups = root
filter_users = root
reconnection_retries = 3

[pam]
reconnection_retries = 3

[domain/foo.bar]
min_id = 1000
max_id = 1000000
use_fully_qualified_names = false
; Using enumerate = true leads to high load and slow response
enumerate = false
cache_credentials = true

id_provider = ldap
auth_provider = ldap
chpass_provider = ldap

ldap_uri = ldap://ldap.foo.bar,ldap://ldap1.foo.bar
ldap_search_base = o=foo-bar
ldap_tls_reqcert = demand
ldap_tls_cacert = /etc/ssl/certs/ca-certificates.crt
ldap_schema = rfc2307bis
ldap_user_search_base = ou=users,o=foo-bar
ldap_user_object_class = eeuser
ldap_group_member = uniquemember
ldap_access_filter = o=foo-bar