We have installed DRBD 8.3.7, and have created an apparently working 3 node replication:<br><br>[root@sweng65 ~]# cat /proc/drbd<br>version: 8.3.7 (api:88/proto:86-91)<br>srcversion: EE47D8BF18AC166BE219757<br> 0: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r----<br>
ns:1081851 nr:0 dw:33447 dr:2097076 al:13 bm:64 lo:0 pe:0 ua:0 ap:0 ep:1 wo:b oos:0<br> 1: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r----<br> ns:1081740 nr:0 dw:33300 dr:1048477 al:13 bm:64 lo:0 pe:0 ua:0 ap:0 ep:1 wo:b oos:0<br>
<br>On the stacked primary resource we have created an ext3 file system via:<br>'mkfs.ext3 /dev/drbd1' <br><br>Our Primary/secondary resource is on /dev/drbd0.<br><br>To test truck replication from the remote backup mirror, we have stopped drbd on both the primary and secondary servers, then followed the procedure described here <a href="http://www.drbd.org/users-guide-legacy/s-using-truck-based-replication.html">http://www.drbd.org/users-guide-legacy/s-using-truck-based-replication.html</a> , using dd to create a bit image copy to a file that we then dd onto the backing store of the (down) primary server's backing store.<br>
<br>Our attempts to get the primary mirror up look like this (athough the ext3 file system is recovered, most importantly at the moment for us is that the stacked metadata is missing, and cannot be re-created without shrinking the ext3 file system ... why would that be true?):<br>
<br>[root@sweng65 ~]# dd if=/pkg/nvs/cores/truck.1008 of=/dev/VG00/mysql bs=256k<br>4096+0 records in<br>4096+0 records out<br>1073741824 bytes (1.1 GB) copied, 6.6437 seconds, 162 MB/s<br>[root@sweng65 ~]# drbdadm --stacked up r0-U<br>
Can not open device '/dev/drbd0': Read-only file system<br>Command 'drbdsetup 1 disk /dev/drbd0 /dev/drbd0 internal --set-defaults --create-device --on-io-error=call-local-io-error' terminated with exit code 20<br>
drbdadm attach r0-U: exited with code 20<br>[root@sweng65 ~]# drbdadm up r0<br>[root@sweng65 ~]# drbdadm --stacked up r0-U<br>Can not open device '/dev/drbd0': Read-only file system<br>Command 'drbdsetup 1 disk /dev/drbd0 /dev/drbd0 internal --set-defaults --create-device --on-io-error=call-local-io-error' terminated with exit code 20<br>
drbdadm attach r0-U: exited with code 20<br>[root@sweng65 ~]# cat /proc/drbd<br>version: 8.3.7 (api:88/proto:86-91)<br>srcversion: EE47D8BF18AC166BE219757<br> 0: cs:WFConnection ro:Secondary/Unknown ds:UpToDate/DUnknown C r----<br>
ns:0 nr:0 dw:0 dr:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:b oos:0<br> 1: cs:Unconfigured<br>[root@sweng65 ~]# drbdadm primary r0<br>[root@sweng65 ~]# drbdadm --stacked up r0-U<br>1: Failure: (119) No valid meta-data signature found.<br>
<br> ==> Use 'drbdadm create-md res' to initialize meta-data area. <==<br><br>Command 'drbdsetup 1 disk /dev/drbd0 /dev/drbd0 internal --set-defaults --create-device --on-io-error=call-local-io-error' terminated with exit code 10<br>
[root@sweng65 ~]# drbdadm --stacked create-md r0-U<br>md_offset 1073598464<br>al_offset 1073565696<br>bm_offset 1073532928<br><br>Found ext3 filesystem<br> 1048440 kB data area apparently used<br> 1048372 kB left usable by current configuration<br>
<br>Device size would be truncated, which<br>would corrupt data and result in<br>'access beyond end of device' errors.<br>You need to either<br> * use external meta data (recommended)<br> * shrink that filesystem first<br>
* zero out the device (destroy the filesystem)<br>Operation refused.<br><br>Command 'drbdmeta 1 v08 /dev/drbd0 internal create-md' terminated with exit code 40<br>drbdadm create-md r0-U: exited with code 40<br>
[root@sweng65 ~]# mount -t ext3 /dev/drbd0 /mirror<br>[root@sweng65 ~]# ls -l /mirror<br>total 16<br>drwx------ 2 root root 16384 Aug 4 15:03 lost+found<br><br><br><br><br><br><br>