88
88
<arg choice="opt" rep="norepeat">-d<arg choice="req" rep="norepeat"><replaceable>hash_alg</replaceable></arg></arg>
89
89
<arg choice="opt" rep="norepeat">-o</arg>
90
90
<arg choice="opt" rep="norepeat">-n</arg>
91
<arg choice="opt" rep="norepeat">-g<arg choice="req" rep="norepeat"><replaceable>congestion_policy</replaceable></arg></arg>
92
<arg choice="opt" rep="norepeat">-f<arg choice="req" rep="norepeat"><replaceable>val</replaceable></arg></arg>
93
<arg choice="opt" rep="norepeat">-h<arg choice="req" rep="norepeat"><replaceable>val</replaceable></arg></arg>
92
95
<cmdsynopsis sepchar=" ">
93
96
<command moreinfo="none">drbdsetup</command>
379
382
<option>--on-io-error <replaceable>err_handler</replaceable></option></term>
381
384
<para> If the driver of the <replaceable>lower_device</replaceable>
382
reports an error to DRBD, DRBD will either pass the error
383
to the upper layers of the operating system, call a helper
385
reports an error to DRBD, DRBD will mark the disk as inconsistent,
384
387
program, or detach the device from its backing storage and
385
388
perform all further IO by requesting it from the peer. The
386
389
valid <replaceable>err_handlers</replaceable> are:
1009
1012
</varlistentry>
1014
<term><option>-g</option>, <option>--on-congestion </option><replaceable>congestion_policy</replaceable></term>
1015
<term><option>-f</option>, <option>--congestion-fill </option><replaceable>fill_threshold</replaceable></term>
1016
<term><option>-h</option>, <option>--congestion-extents </option><replaceable>active_extents_threshold</replaceable></term>
1018
<para>By default DRBD blocks when the available TCP send queue becomes full.
1019
That means it will slow down the application that generates the write
1020
requests that cause DRBD to send more data down that TCP connection.
1022
<para>When DRBD is deployed with DRBD-proxy it might be more desirable that
1023
DRBD goes into AHEAD/BEHIND mode shortly before the send queue becomes full.
1024
In AHEAD/BEHIND mode DRBD does no longer replicate data, but still keeps
1025
the connection open.</para>
1026
<para>The advantage of the AHEAD/BEHIND mode is that the
1027
application is not slowed down, even if DRBD-proxy's buffer is
1028
not sufficient to buffer all write requests. The downside is that
1029
the peer node falls behind, and that a resync will be necessary to
1030
bring it back into sync. During that resync the peer node will have
1031
an inconsistent disk. </para>
1032
<para>Available <replaceable>congestion_policy</replaceable>s are <option>block</option>
1033
and <option>pull-ahead</option>. The default is <option>block</option>.
1034
<replaceable>Fill_threshold</replaceable> might be in the range of 0 to 10GiBytes. The
1035
default is 0 which disables the check. <replaceable>Active_extents_threshold</replaceable>
1036
has the same limits as <option>al-extents</option>.</para>
1037
<para>The AHEAD/BEHIND mode and its settings are available since DRBD 8.3.10.</para>
1010
1040
</variablelist>