Fedora 11 x86_64 with kernel 2.6.30.5-43.fc11.x86_64 and drbd-8.3.3rc1 compiled from source with make rpm<br>so that I have now<br>[root@virtfedbis ]# rpm -qa drbd*<br>drbd-8.3.3rc1-3.x86_64<br>drbd-km-2.6.30.5_43.fc11.x86_64-8.3.3rc1-3.x86_64<br>
<br>The configuration is Primary/Primary<br><br>I get this message on one node<br>Sep 8 17:32:34 virtfedbis kernel: block drbd0: disk( UpToDate -> Failed ) <br>Sep 8 17:32:34 virtfedbis kernel: block drbd0: Local IO failed. Detaching...<br>
Sep 8 17:32:34 virtfedbis kernel: block drbd0: disk( Failed -> Diskless ) <br>Sep 8 17:32:34 virtfedbis kernel: block drbd0: Notified peer that my disk is broken.<br><br>Now "service drdbd status" command on this node gives:<br>
drbd driver loaded OK; device status:<br>version: 8.3.3rc1 (api:88/proto:86-91)<br>GIT-hash: 026d60bb0e6a7d5758c6c3e6245f38f6d8b921aa build by <a href="mailto:root@virtfedbis.ceda.polimi.it">root@virtfedbis.ceda.polimi.it</a>, 2009-09-08 16:21:30<br>
m:res cs ro ds p mounted fstype<br>0:r0 Connected Primary/Primary Diskless/UpToDate C<br><br>2 problems:<br><br>a) it seems I actually didn't get any I/O error in messages, apart from this drbd one.....<br>
how to check if actually I had an I/O error?<br><br>b) how are the proper commands to recover or at least try to recover, supposing the disk is ok?<br><br>The disk is an hw raid on a Hp blade and I don't get any hw error indeed also from information provided by iLO....<br>
Does drdbd support some kind of queuing via drbd.conf, or does it inherit queuing from scsi layer or what else?<br><br>Only messages I get before this event are some minutes before when peer drbd daemon started and so sync happened:<br>
<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: Handshake successful: Agreed network protocol version 91<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: Peer authenticated using 20 bytes of 'sha1' HMAC<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: conn( WFConnection -> WFReportParams ) <br>
Sep 8 17:29:35 virtfedbis kernel: block drbd0: Starting asender thread (from drbd0_receiver [9977])<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: data-integrity-alg: <not-used><br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: drbd_sync_handshake:<br>
Sep 8 17:29:35 virtfedbis kernel: block drbd0: self FFEDAA5E725D8157:0DB564243F5AA9A3:377245292BBD1112:F6DD5DF112448173 bits:0 flags:0<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: peer 0DB564243F5AA9A2:0000000000000000:377245292BBD1113:F6DD5DF112448173 bits:0 flags:0<br>
Sep 8 17:29:35 virtfedbis kernel: block drbd0: uuid_compare()=1 by rule 70<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: peer( Unknown -> Secondary ) conn( WFReportParams -> WFBitMapS ) <br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: peer( Secondary -> Primary ) <br>
Sep 8 17:29:35 virtfedbis kernel: block drbd0: conn( WFBitMapS -> SyncSource ) pdsk( Outdated -> Inconsistent ) <br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: Began resync as SyncSource (will sync 0 KB [0 bits set]).<br>
Sep 8 17:29:35 virtfedbis kernel: block drbd0: Resync done (total 1 sec; paused 0 sec; 0 K/sec)<br>Sep 8 17:29:35 virtfedbis kernel: block drbd0: conn( SyncSource -> Connected ) pdsk( Inconsistent -> UpToDate ) <br>
Sep 8 17:29:40 virtfedbis kernel: block drbd0: md_sync_timer expired! Worker calls drbd_md_sync().<br><br>similar output from dmesg command gives as latest rows:<br><br>block drbd0: drbd_sync_handshake:<br>block drbd0: self FFEDAA5E725D8157:0DB564243F5AA9A3:377245292BBD1112:F6DD5DF112448173 bits:0 flags:0<br>
block drbd0: peer 0DB564243F5AA9A2:0000000000000000:377245292BBD1113:F6DD5DF112448173 bits:0 flags:0<br>block drbd0: uuid_compare()=1 by rule 70<br>block drbd0: peer( Unknown -> Secondary ) conn( WFReportParams -> WFBitMapS ) <br>
block drbd0: peer( Secondary -> Primary ) <br>block drbd0: conn( WFBitMapS -> SyncSource ) pdsk( Outdated -> Inconsistent ) <br>block drbd0: Began resync as SyncSource (will sync 0 KB [0 bits set]).<br>block drbd0: Resync done (total 1 sec; paused 0 sec; 0 K/sec)<br>
block drbd0: conn( SyncSource -> Connected ) pdsk( Inconsistent -> UpToDate ) <br>dlm: connecting to 1<br>block drbd0: md_sync_timer expired! Worker calls drbd_md_sync().<br>block drbd0: disk( UpToDate -> Failed ) <br>
block drbd0: Local IO failed. Detaching...<br>block drbd0: disk( Failed -> Diskless ) <br>block drbd0: Notified peer that my disk is broken.<br><br>Thanks,<br>Gianluca<br><br>