Hi,<br><br>I've been using DRBD for about two
years and since I'm between two freelances contracts, I decided to play
with DRBD a bit more. This mail might be on the long side, but it end
with an help offer so bear with me ;-).<br><br>Install process:<br>I
installed 2 Debian/SID systems in my lab, both of which is a totally
identical, minimal, install. The first system is named Ruby, second one
is Crystal. I installed with the following commands :<br><br># apt-get install git-core<br>
# cd /usr/local/src<br>
# git-clone git://git.drbd.org/drbd-8.2.git drbd-8.2 <br>
# apt-get install linux-headers-`uname -r` build-essential flex docbook-utils<br>
# cd /usr/local/src/drbd-8.2<br>
# make<br>
# make doc<br>
# make install<br><br>I
then modified /etc/drbd.conf to fit my config (/etc/drbd.conf -
available at the url in _1_ ). Everything is working great with the
standard Primary/Secondary mode with an ext3 filesystem.<br><br>Where problems appear :<br><br>I
wanted to try the new "online verifiy". Things then got a bit ugly.
<br><br>Please note : this is no production environment, I'm not freak enough
to put an unstable DRBD (version 8.2.5) on an unstable distribution
(Debian/SID, kernel 2.6.24-1). Those systems are in a lab environment
and available for any test you want me to throw at them. Not sure if
it is a bug or simply a miss-configuration on my side. If it is the
latter feel free to throw me back at my books ;-)<br><br>The status seem good :<br>ruby# cat /proc/drbd<br>version: 8.2.5 (api:88/proto:86-88)<br>GIT-hash: db3f40437ed29d4f4fcefca8154d6972dc101561 build by root@ruby, 2008-02-12 19:45:26<br>
0: cs:Connected st:Primary/Secondary ds:UpToDate/UpToDate C r---<br> ns:4 nr:0 dw:4 dr:17 al:0 bm:0 lo:0 pe:0 ua:0 ap:0<br> resync: used:0/31 hits:0 misses:0 starving:0 dirty:0 changed:0<br> act_log: used:0/257 hits:1 misses:0 starving:0 dirty:0 changed:0<br>
<br>crystal# cat /proc/drbd<br>
version: 8.2.5 (api:88/proto:86-88)
<br>GIT-hash: db3f40437ed29d4f4fcefca8154d6972dc101561 build by root@crystal, 2008-02-12 14:22:58
<br> 0: cs:Connected st:Secondary/Primary ds:UpToDate/UpToDate C r---
<br> ns:0 nr:4 dw:4 dr:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0
<br> resync: used:0/31 hits:0 misses:0 starving:0 dirty:0 changed:0
<br> act_log: used:0/257 hits:0 misses:0 starving:0 dirty:0 changed:0
<br><br>There is a weird warning (on crystal) when I use the mounted fs on ruby : <br>drbd0: local disk flush failed with status -5<br><br>This error does seem to come from ./drbd/drbd_receiver.c concerning BarrierAck, which my dmesg, at drbd startup clearly state:<br>
<pre>drbd0: Barriers not supported on meta data device - disabling<br></pre><br>And, as soon as I use, on Ruby (the primary), this command, it get even weirder:<br><br># drbdadm verify store <br><br>DRBD fail with those errors: <br>
<br>Jade:<br>
<pre>drbd0: conn( Connected -> VerifyS ) <br>drbd0: sock_recvmsg returned -104<br>drbd0: peer( Secondary -> Unknown ) conn( VerifyS -> NetworkFailure ) pdsk( UpToDate -> DUnknown ) <br>drbd0: asender terminated<br>
drbd0: Terminating asender thread<br>drbd0: sock was shut down by peer<br>drbd0: short read expecting header on sock: r=0<br>drbd0: Creating new current UUID<br>drbd0: Writing meta data super block now.<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>
drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: drbd_rs_complete_io() called, but extent not found<br>drbd0: tl_clear()<br>drbd0: Connection closed<br>
drbd0: conn( NetworkFailure -> Unconnected ) <br>drbd0: receiver terminated<br>drbd0: receiver (re)started<br>drbd0: conn( Unconnected -> WFConnection ) <br></pre>(The drbd status is now Primary/unknown, uptodate/Dunknown)<br>
<br>Crystal :<br>
<pre>drbd0: conn( Connected -> VerifyT ) <br>drbd0: unknown command?? on meta m: 0x83740267 c: 8192 l: 6144<br>drbd0: peer( Primary -> Unknown ) conn( VerifyT -> Disconnecting ) pdsk( UpToDate -> DUnknown ) <br>
drbd0: asender terminated<br>drbd0: Terminating asender thread<br>drbd0: Writing meta data super block now.<br>drbd0: short read expecting header on sock: r=-512<br>drbd0: tl_clear()<br>drbd0: Connection closed<br>drbd0: conn( Disconnecting -> StandAlone ) <br>
drbd0: receiver terminated<br>drbd0: Terminating receiver thread<br></pre>(The drbd status is now secondary/unknown, uptodate/Dunknown)<br><br>Some informations you might be interested in : <br>
<pre>#cat /proc/crypto (identical on both sys.)<br>Name : hmac(md5)<br>driver : hmac(md5-generic)<br>module : kernel<br>priority : 0<br>refcnt : 2<br>type : hash<br>blocksize : 64<br>
digestsize : 16<br><br>name : md5<br>driver : md5-generic<br>module : kernel<br>priority : 0<br>refcnt : 3<br>type : digest<br>blocksize : 64<br>digestsize : 16</pre><br>#uname -a (identical on both sys, except for hostname)<br>
Linux ruby 2.6.24-1-686 #1 SMP Mon Feb 11 14:37:45 UTC 2008 i686 GNU/Linux<br><br><br><br>On a side note: <br>As
I've said in the introduction, I'm currently between contract, so I've
got some free time. I wanted to know if there is any interest in 1) an
update to the INSTALL file (with, maybe, INSTALL.debian, INSTALL.suse,
etc...) 2) an interest in documentation translation toward french. Once
again, please note that "no" is a valid answer ;-). <br><br>_1_ : link to googledocs, /etc/drbd.conf : <a href="http://docs.google.com/Doc?id=dhcjg4tt_34hdrpz8dw">http://docs.google.com/Doc?id=dhcjg4tt_34hdrpz8dw</a><br>
<br>Pascal Charest<br><br><br>-- <br>Pascal Charest, Free software consultant (GNU/linux)<br><a href="http://blog.pacharest.com">http://blog.pacharest.com</a>