virtualmin modify-dns --all-domains "bug" on Centos 7 (start-limit systemd)

1 post / 0 new
#1 Mon, 10/05/2015 - 03:43
chrismfz

virtualmin modify-dns --all-domains "bug" on Centos 7 (start-limit systemd)

(And maybe anything using systemd).

I tried a migration from Centos 5 to Centos 7, first of all I ran in the old server a virtualmin modify-dns --all-domains --ttl 600 to adjust TTLs.

It did it of course without errors. Multiple times.

Now the issue:

In the new one, with less zones (~25) I tried the same. (Centos 7). Named crashed. virtualmin modify-dns --all-domains --ttl 600 Updating server domain-here.tld .. Setting default TTL to 600 .. .. done

.. done

And I realized in the end that named wasn't running. Check log:

root@hn /var/log # service named status Redirecting to /bin/systemctl status named.service named.service - Berkeley Internet Name Domain (DNS) Loaded: loaded (/usr/lib/systemd/system/named.service; enabled) Active: failed (Result: start-limit) since Mon 2015-10-05 11:24:36 EEST; 6s ago Process: 6022 ExecStop=/bin/sh -c /usr/sbin/rndc stop > /dev/null 2>&1 || /bin/kill -TERM $MAINPID (code=exited, status=0/SUCCESS) Process: 5962 ExecStart=/usr/sbin/named -u named $OPTIONS (code=exited, status=0/SUCCESS) Process: 5960 ExecStartPre=/usr/sbin/named-checkconf -z /etc/named.conf (code=exited, status=0/SUCCESS) Main PID: 5964 (code=exited, status=0/SUCCESS)

Oct 05 11:24:36 hn.liveevents.gr named[5964]: stopping command channel on 127.0.0.1#953 Oct 05 11:24:36 hn.liveevents.gr named[5964]: no longer listening on ::#53 Oct 05 11:24:36 hn.liveevents.gr named[5964]: no longer listening on 127.0.0.1#53 Oct 05 11:24:36 hn.liveevents.gr named[5964]: no longer listening on 144.76.117.101#53 Oct 05 11:24:36 hn.liveevents.gr named[5964]: exiting Oct 05 11:24:36 hn.liveevents.gr systemd[1]: Stopped Berkeley Internet Name Domain (DNS). Oct 05 11:24:37 hn.liveevents.gr systemd[1]: Starting Berkeley Internet Name Domain (DNS)... Oct 05 11:24:37 hn.liveevents.gr systemd[1]: named.service start request repeated too quickly, refusing to start. Oct 05 11:24:37 hn.liveevents.gr systemd[1]: Failed to start Berkeley Internet Name Domain (DNS). Oct 05 11:24:37 hn.liveevents.gr systemd[1]: Unit named.service entered failed state.

That's a nice one. systemd ftw! failed (Result: start-limit) since Mon 2015-10-05 11:24:36 EEST; 6s ago named.service start request repeated too quickly, refusing to start.

So, with upstart/sysv like init is works just fine (tested again multiple times in Centos 5 and Centos 6). In Centos 7 it crashes (also tested multiple times, everytime it crashed). And of course it leaves named down. (nice one systemd...)

Is there somewhere an option / parameter (and I don't know it yet - fresh to systemd) to raise this ridiculous "start-limit" ?

If not, maybe it's better to make zones updates using rndc reload domain-here.tld ? (but that's virtualmin's functions I suppose I can't change it).

Anyone else can reproduce it on El7 based systems with 25+ zones (that's how many I got and I get the error)