Univention Bugzilla – Bug 32481
remove unnecessary ntpd restarts on interface change
Last modified: 2019-01-03 07:19:04 CET
1. There is be a change in Debian Wheezy in the calling convention of the /etc/network/if-*.d/ scripts: $ man interfaces > When ifupdown is being called with the --all option, ... i[t] calls all the hook scripts ... with IFACE set to "--all", ... diff --git a/branches/ucs-3.2/ucs-3.2-0/base/univention-network-manager/etc/network/ucs-network-tools b/branches/ucs-3.2/ucs-3.2-0/base/univention-network-manager/etc/network/ucs-network-tools index ff70298..ab1e5a1 100644 --- a/branches/ucs-3.2/ucs-3.2-0/base/univention-network-manager/etc/network/ucs-network-tools +++ b/branches/ucs-3.2/ucs-3.2-0/base/univention-network-manager/etc/network/ucs-network-tools @@ -3,7 +3,7 @@ # Univention Network Manager # network script helper functions # -# Copyright 2009-2012 Univention GmbH +# Copyright 2009-2013 Univention GmbH # # http://www.univention.de/ # @@ -31,7 +31,7 @@ # <http://www.gnu.org/licenses/>. ucs_ignore_interface () { - if [ "$IFACE" = "lo" -o "$IFACE" = "all" ]; then + if [ "$IFACE" = "lo" -o "$IFACE" = "--all" ]; then exit 0 fi } 2. On servers with more than one "eth0", (bridging, bond, vlans), a lot of services are restarted multiple times, once for each interface: bind9: should use "rndc reload" instead of a "restart", which is a lot faster and also picks up removed/new interfaces: rm /etc/network/if-*.d/bind9 tail -f /var/log/daemon.log & ifdown eth0 ; ip a flush eth0 ; rndc reload ifup eth0 ; rndc reload Sep 4 15:54:25 bug23670 named[27189]: listening on IPv4 interface eth0, 10.200.17.30#53 ntpd: NTPd picks up new and gone interfaces itself after some time automatically, so no special handling is needed. rm /etc/network/if-*.d/*date Sep 4 15:58:51 bug23670 ntpd[24890]: Deleting interface #14 eth0, fe80::5054:ff:fe52:4145#123, interface stats: received=0, sent=0, dropped=0, active_time=300 secs Sep 4 15:58:51 bug23670 ntpd[24890]: Deleting interface #13 eth0, 10.200.17.30#123, interface stats: received=1, sent=1, dropped=4, active_time=300 secs Sep 4 15:58:51 bug23670 ntpd[24890]: 192.168.0.240 interface 10.200.17.30 -> (null) ... Sep 4 16:03:51 bug23670 ntpd[24890]: Listen normally on 15 eth0 10.200.17.30 UDP 123 Sep 4 16:03:51 bug23670 ntpd[24890]: Listen normally on 16 eth0 fe80::5054:ff:fe52:4145 UDP 123 listener, notifier: TBD...
(In reply to Philipp Hahn from comment #1) > ntpd: > NTPd picks up new and gone interfaces itself after some time > automatically, so no special handling is needed. Indeed this seems to work pretty well. As such I suggest removing /etc/network/if-down.d/00_rdate /etc/network/if-up.d/05_rdate because those unnecessary restarts only hinder NTPd. Also /etc/network/if-up.d/ntpdate should be removed because ntpdate is not useful with NTPd. (Or maybe changed to only act when NTPd is not installed/in use)
(In reply to Janek Walkenhorst from comment #2) > Also > /etc/network/if-up.d/ntpdate > should be removed because ntpdate is not useful with NTPd. This is from the Debian package "ntpdate". > (Or maybe changed > to only act when NTPd is not installed/in use) This is already the case: /usr/sbin/ntpdate-debian -b 19 Sep 12:45:18 ntpdate[32140]: the NTP socket is in use, exiting
*** Bug 32178 has been marked as a duplicate of this bug. ***
This issue has been filled against UCS 4.1. The maintenance with bug and security fixes for UCS 4.1 has ended on 5st of April 2018. Customers still on UCS 4.1 are encouraged to update to UCS 4.3. Please contact your partner or Univention for any questions. If this issue still occurs in newer UCS versions, please use "Clone this bug" or simply reopen the issue. In this case please provide detailed information on how this issue is affecting you.