<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<blockquote cite="mid:5316F0DA.7050704@darktech.org.uk" type="cite">Hi
Matthieu,<br>
<br>
<div class="moz-cite-prefix">On 05/03/2014 07:29, Matthieu Lejeune
wrote:<br>
</div>
<blockquote cite="mid:5316D24E.8040201@exxoss.com" type="cite">
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
Hi all,<br>
<br>
I had a problem this night with a DRBD Primary/Slave.<br>
<br>
<br>
The slave experienced a hardware issue (LSI controller freezed).<br>
It seems the master hold I/O waiting for the slave to respond
until timeout.<br>
<br>
<br>
This caused all targets exported trough infiniband to be
disconnected from the master.<br>
<br>
<br>
So, practically, the master stop responding due to a failure on
the slave.<br>
<br>
I had to hard reboot (power cycle) the slave because UDEV wasn't
responding and did not allow normal reboot.<br>
After slave reboot, drdb did reconnect. It was in status pri/sec
uptodate/uptodate.<br>
But the LSI controller immediatly timeout causing the same issue
a second time.<br>
<br>
<br>
How can we prevent issue on the slave to impact the master ?<br>
<br>
</blockquote>
have a look at ko-count <br>
<br>
<dt style="color: rgb(0, 0, 0); font-family: sans-serif;
font-size: 13px; font-style: normal; font-variant: normal;
font-weight: normal; letter-spacing: normal; line-height:
normal; orphans: auto; text-align: start; text-indent: 0px;
text-transform: none; white-space: normal; widows: auto;
word-spacing: 0px; -webkit-text-stroke-width: 0px;"><span
class="term"><code class="option">ko-count<span
class="Apple-converted-space"> </span><em
class="replaceable"><code>number</code></em></code></span></dt>
<dd style="color: rgb(0, 0, 0); font-family: sans-serif;
font-size: 13px; font-style: normal; font-variant: normal;
font-weight: normal; letter-spacing: normal; line-height:
normal; orphans: auto; text-align: start; text-indent: 0px;
text-transform: none; white-space: normal; widows: auto;
word-spacing: 0px; -webkit-text-stroke-width: 0px;">
<p>In case the secondary node fails to complete a single write
request for<span class="Apple-converted-space"> </span><em
class="replaceable"><code>count</code></em><span
class="Apple-converted-space"> </span>times the<span
class="Apple-converted-space"> </span><em
class="replaceable"><code>timeout</code></em>, it is
expelled from the cluster. (I.e. the primary node goes into<span
class="Apple-converted-space"> </span><code class="option">StandAlone</code><span
class="Apple-converted-space"> </span>mode.) The default
value is 0, which disables this feature.</p>
</dd>
<dt><br>
</dt>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://www.drbd.org/users-guide/re-drbdconf.html">http://www.drbd.org/users-guide/re-drbdconf.html</a><br>
<br>
<blockquote cite="mid:5316D24E.8040201@exxoss.com" type="cite"> <br>
Thank you.<br>
Matthieu Lejeune<br>
<br>
<br>
drbd8-utils : 2:8.3.13-2
amd64 RAID 1 over tcp/ip for Linux utilities<br>
Debian : <br>
root@ifprdstor8a:~/trunk# cat /proc/version <br>
Linux version 3.2.0-4-amd64 (<a moz-do-not-send="true"
class="moz-txt-link-abbreviated"
href="mailto:debian-kernel@lists.debian.org">debian-kernel@lists.debian.org</a>)
(gcc version 4.6.3 (Debian 4.6.3-14) ) #1 SMP Debian 3.2.51-1<br>
root@ifprdstor8a:~/trunk# <br>
<br>
We are using the scst/srpt with the Trunk version of the 7
January 2014<br>
<br>
I give you the config :<br>
<b>drbd global : </b><b><br>
</b><br>
<a moz-do-not-send="true" class="moz-txt-link-abbreviated"
href="mailto:root@ifprdstor8a:/etc/drbd.d#">root@ifprdstor8a:/etc/drbd.d#</a>
cat global_common.conf <br>
global {<br>
usage-count yes;<br>
# minor-count dialog-refresh disable-ip-verification<br>
}<br>
<br>
common {<br>
protocol C;<br>
<br>
handlers {<br>
# The following 3 handlers were disabled due to #576511.<br>
# Please check the DRBD manual and enable them, if they
make sense in your setup.<br>
# pri-on-incon-degr
"/usr/lib/drbd/notify-pri-on-incon-degr.sh;
/usr/lib/drbd/notify-emergency-reboot.sh; echo b >
/proc/sysrq-trigger ; reboot -f";<br>
# pri-lost-after-sb
"/usr/lib/drbd/notify-pri-lost-after-sb.sh;
/usr/lib/drbd/notify-emergency-reboot.sh; echo b >
/proc/sysrq-trigger ; reboot -f";<br>
# local-io-error "/usr/lib/drbd/notify-io-error.sh;
/usr/lib/drbd/notify-emergency-shutdown.sh; echo o >
/proc/sysrq-trigger ; halt -f";<br>
<br>
# fence-peer "/usr/lib/drbd/crm-fence-peer.sh";<br>
# split-brain "/usr/lib/drbd/notify-split-brain.sh
root";<br>
# out-of-sync "/usr/lib/drbd/notify-out-of-sync.sh
root";<br>
# before-resync-target
"/usr/lib/drbd/snapshot-resync-target-lvm.sh -p 15 -- -c 16k";<br>
# after-resync-target
/usr/lib/drbd/unsnapshot-resync-target-lvm.sh;<br>
}<br>
<br>
startup {<br>
# wfc-timeout degr-wfc-timeout outdated-wfc-timeout
wait-after-sb<br>
}<br>
<br>
disk {<br>
# on-io-error fencing use-bmbv no-disk-barrier
no-disk-flushes<br>
# no-disk-drain no-md-flushes max-bio-bvecs<br>
}<br>
<br>
net {<br>
# sndbuf-size rcvbuf-size timeout connect-int ping-int
ping-timeout max-buffers<br>
# max-epoch-size ko-count allow-two-primaries
cram-hmac-alg shared-secret<br>
# after-sb-0pri after-sb-1pri after-sb-2pri
data-integrity-alg no-tcp-cork<br>
}<br>
<br>
syncer {<br>
# rate after al-extents use-rle cpu-mask verify-alg
csums-alg<br>
}<br>
}<br>
<br>
<b>Ressource Configuration : </b><br>
<br>
<a moz-do-not-send="true" class="moz-txt-link-abbreviated"
href="mailto:root@ifprdstor8a:/etc/drbd.d#">root@ifprdstor8a:/etc/drbd.d#</a>
cat DSA801.res <br>
resource DSA801 {<br>
protocol C;<br>
<br>
startup {<br>
wfc-timeout 0;<br>
}<br>
<br>
disk {<br>
on-io-error detach;<br>
}<br>
<br>
syncer {<br>
rate 400M;<br>
verify-alg md5;<br>
}<br>
<br>
on ifprdstor8a {<br>
device /dev/drbd1;<br>
disk /dev/sda;<br>
address 10.13.1.5:7788;<br>
meta-disk internal;<br>
}<br>
<br>
on ifprdstor8b {<br>
device /dev/drbd1;<br>
disk /dev/sda;<br>
address 10.13.1.6:7788;<br>
meta-disk internal;<br>
}<br>
}<br>
<br>
</blockquote>
</blockquote>
<br>
</body>
</html>