<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Hi Janelle,<br>
    <div class="moz-cite-prefix">On 09/01/2015 06:17 PM, Janelle wrote:<br>
    </div>
    <blockquote cite="mid:55E5CF87.9030701@gmail.com" type="cite">
      <meta content="text/html; charset=ISO-8859-1"
        http-equiv="Content-Type">
      On 8/28/15 8:17 AM, Vaclav Adamec wrote:<br>
      <blockquote
cite="mid:CAN1zQ4YRucN_sYziAW7GJFLjHj1RW35BcyUyuDBPfZZap5z5wQ@mail.gmail.com"
        type="cite">
        <div dir="ltr">You could try this (RH recommended way). It works
          for me better than <a moz-do-not-send="true"
            href="http://cleanallruv.pl/" target="_blank"
            style="font-family:verdana,sans-serif;font-size:12.8px">cleanallruv.pl</a><span
            style="font-family:verdana,sans-serif;font-size:12.8px"> as
            this sometimes leads to ldap freeze</span>)
          <div><br>
          </div>
          <div><span
              style="color:rgb(51,51,51);font-family:Overpass,'Open
Sans',Helvetica,sans-serif;font-size:14px;line-height:19.6px;white-space:pre-wrap">unable
              to decode: {replica 30} 5548fa200000001e0000
              5548fa200000001e0000
              unable to decode: {replica 26} 5548a9a80000001a0000
              5548a9a80000001a0000</span><br>
          </div>
          <div><br>
          </div>
          <div>for all of them, on-by-one:</div>
          <div><br>
          </div>
          <div><span
              style="color:rgb(51,51,51);font-family:Overpass,'Open
Sans',Helvetica,sans-serif;font-size:14px;line-height:19.6px;white-space:pre-wrap">ldapmodify
              -x -D "cn=directory manager" -w XXXXXXX
              dn: cn=replica,cn=dc\3Dexample\2Cdc\3Dcom,cn=mapping
              tree,cn=config
              changetype: modify
              replace: nsds5task
              nsds5task: CLEANRUV30
              <enter> + <Ctrl + D></span><br>
          </div>
        </div>
        <div class="gmail_extra"><br>
          <div class="gmail_quote">On Fri, Aug 28, 2015 at 4:55 PM,
            Guillermo Fuentes <span dir="ltr"><<a
                moz-do-not-send="true" class="moz-txt-link-abbreviated"
                href="mailto:guillermo.fuentes@modernizingmedicine.com">guillermo.fuentes@modernizingmedicine.com</a>></span>
            wrote:<br>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div dir="ltr">
                <div class="gmail_default">
                  <div class="gmail_default"><font face="verdana,
                      sans-serif">Hi Janelle,</font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif"><br>
                    </font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif">Using the <a moz-do-not-send="true"
                        href="http://cleanallruv.pl" target="_blank">cleanallruv.pl</a>
                      tool was the only way I was able to get ride of
                      the "</font>unable to decode: {replica x}" entries<span
                      style="font-family:verdana,sans-serif">.</span></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif"><br>
                    </font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif">This is how I used it, cleaning a
                      replica ID at a time:</font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif"># For replica id: 40</font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif"><a moz-do-not-send="true"
                        href="http://cleanallruv.pl" target="_blank">cleanallruv.pl</a>
                      -v -D "cn=directory manager" -w - -b
                      'dc=example,dc=com' -r 40 </font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif"><br>
                    </font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif">Note that the "-w -"​ will make the
                      tool prompt you for the directory manager
                      password.</font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif"><br>
                    </font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif">Hope this helps,</font></div>
                  <div class="gmail_default"><font face="verdana,
                      sans-serif">Guillermo​</font></div>
                  <div class="gmail_default"><br>
                  </div>
                </div>
                <div class="gmail_extra"><br>
                  <div class="gmail_quote">
                    <div>
                      <div class="h5">On Thu, Aug 27, 2015 at 10:27 AM,
                        Janelle <span dir="ltr"><<a
                            moz-do-not-send="true"
                            class="moz-txt-link-abbreviated"
                            href="mailto:janellenicole80@gmail.com">janellenicole80@gmail.com</a>></span>
                        wrote:<br>
                      </div>
                    </div>
                    <blockquote class="gmail_quote" style="margin:0px
                      0px 0px
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
                      <div>
                        <div class="h5">
                          <div bgcolor="#FFFFFF" text="#000000">
                            <div>
                              <div> On 8/27/15 1:05 AM, thierry bordaz
                                wrote:<br>
                                <blockquote type="cite">
                                  <div>On 08/27/2015 09:41 AM, Ludwig
                                    Krispenz wrote:<br>
                                  </div>
                                  <blockquote type="cite"> <br>
                                    On 08/27/2015 09:08 AM, Martin Kosek
                                    wrote: <br>
                                    <blockquote type="cite">On
                                      08/26/2015 05:31 PM, Simo Sorce
                                      wrote: <br>
                                      <blockquote type="cite">On Wed,
                                        2015-08-26 at 06:36 -0700,
                                        Janelle wrote: <br>
                                        <blockquote type="cite">Hello
                                          all, <br>
                                          <br>
                                          My biggest problem is losing
                                          replicas and then trying to
                                          delete the <br>
                                          entries and rebuild them. Here
                                          is a perfect example, I simply
                                          can't get <br>
                                          rid of these  (see below). I
                                          have tried (of course after
                                          the ORIGINAL <br>
                                          "ipa-replica-manage del
                                          hostname --force --clean": <br>
                                          <br>
                                          ipa-replica-manage clean-ruv
                                          25 <br>
                                          <br>
                                          ldapmodify... with: <br>
                                              dn: cn=clean 25,
                                          cn=cleanallruv, cn=tasks,
                                          cn=config <br>
                                              objectclass:
                                          extensibleObject <br>
                                              replica-base-dn:
                                          dc=example,dc=com <br>
                                              replica-id: 25 <br>
                                              cn: clean 25 <br>
                                          <br>
                                          And yet nothing works. Any
                                          suggestions? This is perhaps
                                          the most <br>
                                          frustrating part about
                                          maintaining IPA. <br>
                                          <br>
                                          ~J <br>
                                          <br>
                                          unable to decode: {replica 12}
                                          5588dc2e0000000c0000
                                          559f3de60004000c0000 <br>
                                          unable to decode: {replica 14}
                                          5587aa8d0000000e0000
                                          5587aa8d0003000e0000 <br>
                                          unable to decode: {replica 16}
                                          5588f58f000000100000
                                          55bb7b08000500100000 <br>
                                          unable to decode: {replica 25}
                                          55a4887b000000190000
                                          55a49242000400190000 <br>
                                          unable to decode: {replica 29}
                                          55d199a50001001d0000
                                          55d199a50001001d0000 <br>
                                          unable to decode: {replica 3}
                                          5587c5c3000000030000
                                          55b8a049000100030000 <br>
                                          unable to decode: {replica 5}
                                          55cc82ab041d00050000
                                          55cc82ab041d00050000 <br>
                                        </blockquote>
                                        Have you tried restarting DS
                                        before trying to clean the ruv ?
                                        <br>
                                        <br>
                                        I run in a similar problem in a
                                        test install recently, and I got
                                        better <br>
                                        results that way. The bug is
                                        known to the DS people and they
                                        are working <br>
                                        to get out patches that fix the
                                        root issue. <br>
                                        <br>
                                        Simo. <br>
                                      </blockquote>
                                      CCing DS folks. Wasn't there a
                                      recent DS fix that was supposed to
                                      improve the <br>
                                      RUV situation? <br>
                                      <br>
                                      Looking at 389 DS Trac, I see some
                                      interesting RUV fixes in 1.3.4.x
                                      releases: <br>
                                      <br>
                                      <a moz-do-not-send="true"
href="https://fedorahosted.org/389/query?summary=%7ERUV&status=closed&order=milestone&col=id&col=summary&col=status&col=owner&col=type&col=priority&col=milestone"
                                        target="_blank">https://fedorahosted.org/389/query?summary=~RUV&status=closed&order=milestone&col=id&col=summary&col=status&col=owner&col=type&col=priority&col=milestone</a>
                                      <br>
                                      <br>
                                      I see that 389-ds-base-1.3.4.3 is
                                      already in Fedora 22+, does the
                                      RUV issue <br>
                                      happen there? <br>
                                    </blockquote>
                                    it should not, and I think Thierry
                                    verified the fix. <br>
                                    The problem we resolved and which we
                                    think is the core of the corrupted
                                    RUV was that the cleanallruv task
                                    did only purge the RUV, but dit not
                                    purge the changelog. If cleanallruv
                                    was run and the server had a
                                    disorderly shutdown (crash or abort
                                    when shutdown was hanging) then at
                                    restart the changelog RUV was
                                    rebuilt from the data in the
                                    changelog and if it contained a csn
                                    from cleaned RIDs this was added to
                                    the RUV (but the reference to the
                                    server was lost and so the url part
                                    is missing from this RUV. <br>
                                    The fix now does remove all
                                    references to the cleaned RID from
                                    the changelog and the problem should
                                    not reoccur with RIDs cleaned with
                                    the fix, of course th echangelog can
                                    still can contain references to RIDs
                                    cleaned before the fix - and if no
                                    changelog trimming is configured
                                    this is what will happen. So, even
                                    after the fix old RUVs could pop up
                                    and have to be (finally) cleaned. <br>
                                    <br>
                                    The other source is that these
                                    corrupted rivs can be "imported"
                                    from another server by exchanging
                                    ruvs in the repl protocol.
                                    Cleanallruv tries to address this
                                    and to propagate the cleanallruv
                                    tasks to all servers it thinks are
                                    connected. If there are replication
                                    agreements to servers which no
                                    longer exist or to servers which
                                    cannot be connetcted this will delay
                                    the ruv cleaning <br>
                                    <br>
                                  </blockquote>
                                  <br>
                                  <font face="Times New Roman, Times,
                                    serif">Hello, <br>
                                    <br>
                                    I verified the fix in 1.3.4.2 F22 /
                                    389-ds-base-1.3.4.0-6.el7 RHEL7, so
                                    after those versions CLEANALLRUV do
                                    not create any longer corrupted ruv
                                    elements.<br>
                                    According to the timestamp in the
                                    ruv (for example csn2date.py
                                    5587aa8d0003000e0000 -->
                                    22/06/2015:06:26:21) this are old
                                    ruv elements. I think Ludwig is
                                    right, these corrupted ruv-elements
                                    come from old cleanallruv before the
                                    fix was applied.<br>
                                    <br>
                                    The problem is that even a fixed
                                    server can get those corrupted
                                    ruv-elements from others servers.<br>
                                    All servers in the topology should
                                    be updated with that fix, so that at
                                    least they stop creating corrupted
                                    ruv-elements.<br>
                                    Now to get rid of the existing ones,
                                    I imagine only brute option of
                                    recreating replica and reinit... I
                                    hope an other option is possible.<br>
                                    <br>
                                    thanks<br>
                                    thierry<br>
                                  </font><br>
                                </blockquote>
                              </div>
                            </div>
                          </div>
                        </div>
                      </div>
                    </blockquote>
                  </div>
                </div>
              </div>
            </blockquote>
          </div>
        </div>
      </blockquote>
      <font color="#888888">For a few minutes - almost an hour
        actually,  I thought there was hope.  I found the cleanallruv.pl
        script and that not only seemed to work, but it wiped the
        "unable to decode" from all the servers even just running it on
        one.  Sadly, within an hour, they all came back. :-(<br>
        <br>
        unable to decode  {replica 12} 559f3de60004000c0000
        559f3de60004000c0000<br>
        unable to decode  {replica 14} 5587aa8d0003000e0000
        5587aa8d0003000e0000<br>
        unable to decode  {replica 16} 55bb7b08000500100000
        55bb7b08000500100000<br>
        unable to decode  {replica 25} 55a49242000400190000
        55a49242000400190000<br>
        unable to decode  {replica 29} 55d199a50001001d0000
        55d199a50001001d0000<br>
        unable to decode  {replica 31} 55e4bc680005001f0000
        55e4bc680005001f0000<br>
        unable to decode  {replica 3} 55b8a049000100030000
        55b8a049000100030000<br>
        unable to decode  {replica 5} 55cc82ab041d00050000
        55cc82ab041d00050000<br>
        <br>
        I cried... Followed by heavy drinking.<br>
      </font></blockquote>
    <font color="#888888">does drinking  help, could be  a great
      workaround ?<br>
      <br>
      Now, more seriously, I think you need a build including the
      mentioned improvement for cleanallruv, we are currently checking
      if and where it is available for 7.1.<br>
      But this fix will only help in future cleanallruvs, so you
      probably need to go thru a few iterations of cleaning.<br>
      Since the core problem of the corrupted ruvs is that they can be
      recreated from the changlog I think configuring changlog trimming
      is something that should be done.<br>
    </font>
    <blockquote cite="mid:55E5CF87.9030701@gmail.com" type="cite"><font
        color="#888888"> <br>
        ~Janelle<br>
      </font> <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
    </blockquote>
    <br>
  </body>
</html>