<div dir="ltr"><p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Hey,.. well I know many
shortcuts ;-P SLES, RHEL aka EL,etc... so XENserver &quot;must&quot; be fine
too.</span><span style="font-size:12pt;font-family:&#39;Times New Roman&#39;,serif"></span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif"> </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">My goal is to have:</span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">a Server for KVM hosting... Mail server (Open
change + Cyrus + postfix,...) Fileserver, AD DC using Samba 4, </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">and some more services...</span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif"> </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">If you Ask yourself why I try to do this using an
Atom 2750 8 Core ad 16 GB on Standard Sata disks I can answer it easily. I have
to serve this services for 5 User and I do not have to care about High IO loads
etc, to separate sequential from random IO to push the performance higher.</span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif"> </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">My second Problem is that this server have their
place in very small room without any possibility of cooling. So Atom with 20W
TDP each must do this. </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif"> </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">another reason is the physical size,.. </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">If I would try to put one of my HP DL 380 I have at
home, I would not be able to close the door..</span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif"> </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">So better don&#39;t ask why I choose this components.
It even took me 2 hours to disassemble and assemble the rack again in fact </span><span style="font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">the door frame was 3mm smaller :-P</span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-family:Arial,sans-serif;background-image:initial;background-repeat:initial"> </span></p>

<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:12pt;font-family:Arial,sans-serif">BUT I must see the good things: SCSI Cluster no
longer exists…. My first and my worst shared storage.</span></p><div><br></div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">2015-02-09 17:42 GMT+01:00 Digimer <span dir="ltr">&lt;<a href="mailto:lists@alteeve.ca" target="_blank">lists@alteeve.ca</a>&gt;</span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Sorry, &quot;EL&quot; is a short way of saying &quot;RHEL, CentOS or other RHEL-based<br>
distroes&quot;. CentOS 6 is fine.<br>
<br>
Please don&#39;t let yourself get too frustrated. HA clustering is not hard,<br>
but it is complex. It takes time to get it all working right.<br>
<br>
What is your primary goal? To run an HA cluster for hosting VMs?<br>
<br>
digimer<br>
<span class=""><br>
On 09/02/15 11:32 AM, Patrick Prilisauer wrote:<br>
&gt; Hey, I don&#39;t own EL but CentOS6 must do it also?.<br>
&gt;<br>
&gt; I know sitting for at least two weeks on setting up the Server. I&#39;m<br>
&gt; getting really frustrated.<br>
&gt;<br>
&gt; I&#39;ve bought the combination of Supermicro and Adaptec to do not have to<br>
&gt; handle mdraid + uefi + workaround...<br>
&gt;<br>
&gt; I will try to do the very last setup on EL6.. supermicro was my worst<br>
&gt; decision...<br>
&gt;<br>
&gt; I&#39;ll be back ;-)<br>
&gt;<br>
&gt; 2015-02-09 17:10 GMT+01:00 Digimer &lt;<a href="mailto:lists@alteeve.ca">lists@alteeve.ca</a><br>
</span>&gt; &lt;mailto:<a href="mailto:lists@alteeve.ca">lists@alteeve.ca</a>&gt;&gt;:<br>
<div><div class="h5">&gt;<br>
&gt;     Hello,<br>
&gt;<br>
&gt;       Trying to adapt <a href="https://alteeve.ca/w/AN!Cluster_Tutorial_2" target="_blank">https://alteeve.ca/w/AN!Cluster_Tutorial_2</a> to EL7 +<br>
&gt;     pacemaker will be no easy task at all. The concepts can port, but after<br>
&gt;     that, you&#39;re pretty much on your own.<br>
&gt;<br>
&gt;       Out of curiosity, can you try pacemaker 1.1.12 on cman+corosync on<br>
&gt;     EL6? I&#39;ve avoided EL7 so far because it&#39;s still very very new, and a<br>
&gt;     major departure from previous releases, so I am not convinced it&#39;s ideal<br>
&gt;     for HA yet.<br>
&gt;<br>
&gt;       If you find that the RAID stuff works better on EL6, then it might be<br>
&gt;     a sign of driver issues or something. Truth be told though, my early<br>
&gt;     experiments with RAID and Supermicro didn&#39;t end well, so this might well<br>
&gt;     be not OS related, too. Any firmware updates available?<br>
&gt;<br>
&gt;     On 09/02/15 11:02 AM, Patrick Prilisauer wrote:<br>
&gt;     &gt; Hey there,<br>
&gt;     &gt;<br>
&gt;     &gt; Ih have found out one issue, the Problem was or better still is a<br>
&gt;     &gt; defunced mainboard.<br>
&gt;     &gt; I&#39;m running an Adaptec 6504e on a Supermicro A1ASI-2750 the Mainboard<br>
&gt;     &gt; occures some strange timeouts, which hasn&#39;t been reported by the aacraid<br>
&gt;     &gt; driver because they were to short.<br>
&gt;     &gt; But suddenly after a few days the timeouts getting larger.. and some<br>
&gt;     &gt; other strange things happened linke the uefi bios doesn&#39;t detect the<br>
&gt;     &gt; card etc.<br>
&gt;     &gt;<br>
&gt;     &gt; Anyway, I&#39;m still fighting with the pacemakersetup, I do not understand<br>
&gt;     &gt; the drbd documentation, the configs shown on one chapter doesn&#39;t match<br>
&gt;     &gt; the resources in another chapter...<br>
&gt;     &gt;<br>
&gt;     &gt; Still if somebody knows where I could find a complete configuration made<br>
&gt;     &gt; with the pcs programm for pacemaker. I would be so happy.<br>
&gt;     &gt;<br>
&gt;     &gt; BR<br>
&gt;     &gt;<br>
&gt;     &gt; 2015-02-04 10:07 GMT+01:00 Patrick Prilisauer &lt;<a href="mailto:prilisauer@googlemail.com">prilisauer@googlemail.com</a> &lt;mailto:<a href="mailto:prilisauer@googlemail.com">prilisauer@googlemail.com</a>&gt;<br>
</div></div>&gt;     &gt; &lt;mailto:<a href="mailto:prilisauer@googlemail.com">prilisauer@googlemail.com</a><br>
<div><div class="h5">&gt;     &lt;mailto:<a href="mailto:prilisauer@googlemail.com">prilisauer@googlemail.com</a>&gt;&gt;&gt;:<br>
&gt;     &gt;<br>
&gt;     &gt;     Hello to all,<br>
&gt;     &gt;<br>
&gt;     &gt;     I have now completely reinstalled my servers<br>
&gt;     &gt;     according <a href="https://alteeve.ca/w/AN!Cluster_Tutorial_2" target="_blank">https://alteeve.ca/w/AN!Cluster_Tutorial_2</a> on Centos 7<br>
&gt;     &gt;     modified by some OS differences.<br>
&gt;     &gt;     Anyway, it seems that I won&#39;t understand what I&#39;m doing wrong?<br>
&gt;     &gt;     As you could see res_drbd_1_stop on.... &#39;not installed&#39; is<br>
&gt;     returned.<br>
&gt;     &gt;     The last whole day spending on findinding my needle in the<br>
&gt;     haystack<br>
&gt;     &gt;     wont make it better.<br>
&gt;     &gt;<br>
&gt;     &gt;     Any hints?,<br>
&gt;     &gt;     Thanks<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;     OUTPUT1:<br>
&gt;     &gt;     crm_mon<br>
&gt;     &gt;     Last updated: Wed Feb  4 09:59:01 2015<br>
&gt;     &gt;     Last change: Wed Feb  4 05:18:41 2015 via crmd on at01srv01<br>
&gt;     &gt;     Stack: corosync<br>
&gt;     &gt;     Current DC: at01srv02 (167772162) - partition with quorum<br>
&gt;     &gt;     Version: 1.1.10-32.el7_0.1-368c726<br>
&gt;     &gt;     2 Nodes configured<br>
&gt;     &gt;     4 Resources configured<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;     Node at01srv01 (167772161): standby<br>
&gt;     &gt;     Online: [ at01srv02 ]<br>
&gt;     &gt;<br>
&gt;     &gt;      Master/Slave Set: ms_drbd_1 [res_drbd_1]<br>
&gt;     &gt;          res_drbd_1 (ocf::linbit:drbd):     FAILED at01srv02<br>
&gt;     (unmanaged)<br>
&gt;     &gt;          Stopped: [ at01srv01 ]<br>
&gt;     &gt;     stonith_fence_pcmk_1    (stonith:fence_pcmk):   Started at01srv02<br>
&gt;     &gt;<br>
&gt;     &gt;     Failed actions:<br>
&gt;     &gt;         res_drbd_1_stop_0 on at01srv02 &#39;not installed&#39; (5): call=120,<br>
&gt;     &gt;     status=complete, last-rc-change=&#39;Wed Feb  4 09:3<br>
&gt;     &gt;     5:43 2015&#39;, queued=15118ms, exec=0ms<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;     OUTPUT2:<br>
&gt;     &gt;     [root@at01srv02 linbit]# systemctl status drbd.service<br>
&gt;     &gt;     drbd.service - DRBD -- please disable. Unless you are NOT using a<br>
&gt;     &gt;     cluster manager.<br>
&gt;     &gt;        Loaded: loaded (/usr/lib/systemd/system/drbd.service; disabled)<br>
&gt;     &gt;        Active: failed (Result: exit-code) since Mit 2015-02-04<br>
&gt;     09:57:06<br>
&gt;     &gt;     CET; 2min 51s ago<br>
&gt;     &gt;       Process: 8544 ExecStart=/sbin/drbdadm adjust-with-progress all<br>
&gt;     &gt;     (code=exited, status=1/FAILURE)<br>
&gt;     &gt;       Process: 8540 ExecStartPre=/sbin/drbdadm sh-nop (code=exited,<br>
&gt;     &gt;     status=0/SUCCESS)<br>
&gt;     &gt;      Main PID: 8544 (code=exited, status=1/FAILURE)<br>
&gt;     &gt;<br>
&gt;     &gt;     Feb 04 09:57:06 at01srv02 drbdadm[8544]: [<br>
&gt;     &gt;     Feb 04 09:57:06 at01srv02 drbdadm[8544]: adjust net:<br>
&gt;     &gt;     r0:failed(net-options:20)<br>
&gt;     &gt;     Feb 04 09:57:06 at01srv02 drbdadm[8544]: ]<br>
&gt;     &gt;     Feb 04 09:57:06 at01srv02 systemd[1]: drbd.service: main process<br>
&gt;     &gt;     exited, code=exited, status=1/FAILURE<br>
&gt;     &gt;     Feb 04 09:57:06 at01srv02 systemd[1]: Failed to start DRBD --<br>
&gt;     please<br>
&gt;     &gt;     disable. Unless you are NOT using a c...ger..<br>
&gt;     &gt;     Feb 04 09:57:06 at01srv02 systemd[1]: Unit drbd.service entered<br>
&gt;     &gt;     failed state.<br>
&gt;     &gt;     Hint: Some lines were ellipsized, use -l to show in full.<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;     2015-02-02 20:46 GMT+01:00 Digimer &lt;<a href="mailto:lists@alteeve.ca">lists@alteeve.ca</a><br>
&gt;     &lt;mailto:<a href="mailto:lists@alteeve.ca">lists@alteeve.ca</a>&gt;<br>
</div></div>&gt;     &gt;     &lt;mailto:<a href="mailto:lists@alteeve.ca">lists@alteeve.ca</a> &lt;mailto:<a href="mailto:lists@alteeve.ca">lists@alteeve.ca</a>&gt;&gt;&gt;:<br>
<div><div class="h5">&gt;     &gt;<br>
&gt;     &gt;         On 02/02/15 02:44 PM, Ivan wrote:<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;                     I&#39;m not sure that two (or more) LUKS<br>
&gt;     partitions are<br>
&gt;     &gt;                     identical given<br>
&gt;     &gt;                     exactly the same cleartext content and the same<br>
&gt;     &gt;                     keys. There must be some<br>
&gt;     &gt;                     kind of sector randomization when writing data to<br>
&gt;     &gt;                     make cryptoanalysis<br>
&gt;     &gt;                     harder, so it makes me think that it&#39;s not the<br>
&gt;     case<br>
&gt;     &gt;                     (that would require<br>
&gt;     &gt;                     testing though).<br>
&gt;     &gt;                     If I&#39;m right, I don&#39;t see how DRBD could work in<br>
&gt;     &gt;                     that setup. (or maybe I<br>
&gt;     &gt;                     just need more sleep).<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;                 LUKS is working on the LV, which will be backed by the<br>
&gt;     &gt;                 PV on DRBD. DRBD<br>
&gt;     &gt;                 doesn&#39;t know data, so it will simply replicate the<br>
&gt;     LUKS<br>
&gt;     &gt;                 structure<br>
&gt;     &gt;                 faithfully to both nodes.<br>
&gt;     &gt;<br>
&gt;     &gt;                 Remember, for all intent and purpose, there is<br>
&gt;     only one<br>
&gt;     &gt;                 device/luks<br>
&gt;     &gt;                 partition. DRBD is really no different from LUKS on<br>
&gt;     &gt;                 /dev/mdX devices in<br>
&gt;     &gt;                 this regard.<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;             ah that&#39;s right - indeed more sleep needed. I&#39;ve<br>
&gt;     skipped the<br>
&gt;     &gt;             &quot;clustered<br>
&gt;     &gt;             LVM&quot; part and was thinking about two luks partitions.<br>
&gt;     &gt;<br>
&gt;     &gt;             sorry for the noise.<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;         No worries at all. When you ask a question like this, you<br>
&gt;     have a<br>
&gt;     &gt;         chance to learn a system better, so it&#39;s good. :)<br>
&gt;     &gt;<br>
&gt;     &gt;         --<br>
&gt;     &gt;         Digimer<br>
&gt;     &gt;         Papers and Projects: <a href="https://alteeve.ca/w/" target="_blank">https://alteeve.ca/w/</a><br>
&gt;     &gt;         What if the cure for cancer is trapped in the mind of a person<br>
&gt;     &gt;         without access to education?<br>
&gt;     &gt;         _________________________________________________<br>
&gt;     &gt;         drbd-user mailing list<br>
&gt;     &gt;         <a href="mailto:drbd-user@lists.linbit.com">drbd-user@lists.linbit.com</a> &lt;mailto:<a href="mailto:drbd-user@lists.linbit.com">drbd-user@lists.linbit.com</a>&gt;<br>
</div></div>&gt;     &lt;mailto:<a href="mailto:drbd-user@lists.linbit.com">drbd-user@lists.linbit.com</a> &lt;mailto:<a href="mailto:drbd-user@lists.linbit.com">drbd-user@lists.linbit.com</a>&gt;&gt;<br>
<span class="im HOEnZb">&gt;     &gt;         <a href="http://lists.linbit.com/__mailman/listinfo/drbd-user" target="_blank">http://lists.linbit.com/__mailman/listinfo/drbd-user</a><br>
&gt;     &gt;         &lt;<a href="http://lists.linbit.com/mailman/listinfo/drbd-user" target="_blank">http://lists.linbit.com/mailman/listinfo/drbd-user</a>&gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;     &gt;<br>
&gt;<br>
&gt;<br>
&gt;     --<br>
&gt;     Digimer<br>
&gt;     Papers and Projects: <a href="https://alteeve.ca/w/" target="_blank">https://alteeve.ca/w/</a><br>
&gt;     What if the cure for cancer is trapped in the mind of a person without<br>
&gt;     access to education?<br>
&gt;<br>
&gt;<br>
<br>
<br>
</span><div class="HOEnZb"><div class="h5">--<br>
Digimer<br>
Papers and Projects: <a href="https://alteeve.ca/w/" target="_blank">https://alteeve.ca/w/</a><br>
What if the cure for cancer is trapped in the mind of a person without<br>
access to education?<br>
</div></div></blockquote></div><br></div>