Uploaded image for project: 'Support'
  1. Support
  2. SUPPORT-41

ods hangs after creation new zsk

    XMLWordPrintable

    Details

    • Type: Support
    • Status: Closed
    • Priority: Minor
    • Resolution: Cannot Reproduce
    • Affects Version/s: OpenDNSSEC 1.3.5
    • Fix Version/s: None
    • Component/s: Enforcer, Signer
    • Labels:
      None
    • Environment:

      RedHat 5.8
      ODS 1.3.5, compiled (then packaged to rpm)
      Luna SA
      zonefile .nl
      ca 5000000 domains
      ca 1300000 DS RR

      Description

      On Oct 28th, after creation of a new ZSK, enforcerd and signerd didn't work anymore.
      But ps still showed the processes.

      Oct 27 14:36:43, output ods-ksmutil key list -v
      nl KSK active 2016-12-05 09:38:15 6*****b HSM 21362
      nl ZSK active 2012-10-28 19:36:01 3*****4 HSM 37666

      So on Oct 28th a zsk rollover started:

      Oct 28 12:37:09 signerp2 ods-enforcerd: HSM connection open.
      Oct 28 12:37:09 signerp2 ods-enforcerd: Reading config "/etc/opendnssec/conf.xml"
      Oct 28 12:37:09 signerp2 ods-enforcerd: Reading config schema "/usr/share/opendnssec/conf.rng"
      Oct 28 12:37:09 signerp2 ods-enforcerd: Communication Interval: 3600
      Oct 28 12:37:09 signerp2 ods-enforcerd: No DS Submit command supplied
      Oct 28 12:37:09 signerp2 ods-enforcerd: SQLite database set to: /var/lib/opendnssec/kasp.db
      Oct 28 12:37:09 signerp2 ods-enforcerd: Log User set to: local5
      Oct 28 12:37:09 signerp2 ods-enforcerd: Switched log facility to: local5
      Oct 28 12:37:09 signerp2 ods-enforcerd: Connecting to Database...
      Oct 28 12:37:09 signerp2 ods-enforcerd: Policy default found.
      Oct 28 12:37:09 signerp2 ods-enforcerd: Key sharing is Off.
      Oct 28 12:37:10 signerp2 ods-enforcerd: Created ZSK size: 1024, alg: 8 with id: 2*****d in repository: HSM and database.
      Oct 28 12:37:10 signerp2 ods-enforcerd: zonelist filename set to /etc/opendnssec/zonelist.xml.
      Oct 28 12:37:10 signerp2 ods-enforcerd: Zone nl found.
      Oct 28 12:37:10 signerp2 ods-enforcerd: Policy for nl set to default.
      Oct 28 12:37:10 signerp2 ods-enforcerd: Config will be output to /var/lib/opendnssec/signconf/nl.xml.

      After this, no logging of enforcerd nor signerd.
      I should expect the lines:
      Oct 28 11:37:09 signerp2 ods-enforcerd: Disconnecting from Database...
      Oct 28 11:37:09 signerp2 ods-enforcerd: Sleeping for 3600 seconds.

      ps showed both daemons, but they didn't work.

      Our conf.xml contains:
      <NotifyCommand>/usr/local/zonefile/bin/do_publishsignedzone</NotifyCommand>
      The zonefile was signed, do_publishsignedzone was already started.

      Oct 28 12:28:40 signerp2 do_publishsignedzone: /usr/local/zonefile/bin/do_publishsignedzone started with PID 13136

      Actions do_publishsignezone:
      check of the signed zonefile,
      compile to format raw in the Bind tree,
      rndc reload,
      ods-ksmutil output to syslog

      After completion of do_publish_signedzone ods always writes a STATS line to syslog.

      syslog shows the rndc reload command:
      Oct 28 12:37:10 signerp2 ods-enforcerd: Config will be output to /var/lib/opendnssec/signconf/nl.xml.
      Oct 28 12:37:24 signerp2 do_publishsignedzone: moving zonefile into place
      Oct 28 12:37:25 signerp2 named[5886]: received control channel command 'reload'
      Oct 28 12:37:25 signerp2 named[5886]: loading configuration from '/etc/named/nl/named.conf'
      Oct 28 12:37:25 signerp2 named[5886]: reading built-in trusted keys from file '/etc/bind.keys'
      Oct 28 12:37:25 signerp2 named[5886]: using default UDP/IPv4 port range: [1024, 65535]
      Oct 28 12:37:25 signerp2 named[5886]: using default UDP/IPv6 port range: [1024, 65535]
      Oct 28 12:37:25 signerp2 named[5886]: reloading configuration succeeded
      Oct 28 12:37:25 signerp2 named[5886]: reloading zones succeeded
      Oct 28 12:37:26 signerp2 logger: /etc/init.d/ns-nl: /usr/local/bind/sbin/rndc -c /etc/named/nl/rndc.conf reload
      No further logging, no enforcerd wake-ups.

      And the next zonefile:

      Oct 28 14:15:03 signerp2 do_zonefile: signalling ods to start signing
      no action at all

      Only after restart of ODS things began to work again.

      I have two questions:

      • could this be related to the change in Daylight Saving Time?
        perhaps because of timestamps in kasp.db
      • could there be a relationship with the running script do_publishsignedzone?
        the bind-related actions were completed, but the ods-ksmutil command didn't show up in syslog
        nor the STATS line, so do_publishsignedzone never completed within ODS

        Attachments

          Activity

            People

            Assignee:
            sion SiƓn Lloyd
            Reporter:
            unixbeheer unixbeheer@sidn.nl
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: