Hello everyone,<br><br>I've read all the posts regarding the above situation, but could not get anything to work in our installation.<br><br>After the primary server gets back post forced reboot, the secondary server won't let go off the primary drbd state. Heartbet is working fine and calling the drbddisk within its resources directory.
<br><br>lsof and fuser don't report anything using /dev/drbd0 or /var/lib/mysql (which is what's being synced across machines), but I went even further by trying to disable almost all services and rmmod'ing several modules. Yet, secondary stubbornly refuses to give it away.
<br><br>After reading the other posts, I assume it has something to do with the kernel? Both machines use Centos 4.4, with 2.6.9-42.0.3.ELsmp.<br><br>Secondary server says:<br><br>machine2# cat /proc/drbd<br>version: 0.7.21
(api:79/proto:74)<br>SVN Revision: 2326 build by buildsvn@build-i386, 2006-10-07 05:12:17<br> 0: cs:Connected st:Primary/Secondary ld:Consistent<br> ns:94216 nr:0 dw:40 dr:94614 al:0 bm:29 lo:0 pe:0 ua:0 ap:0<br><br>if I try to force drbd into secondary state it gives me:
<br><br>machine2# drbdadm secondary all<br>ioctl(,SET_STATE,) failed: Device or resource busy<br>Someone has opened the device for RW access!<br>Command 'drbdsetup /dev/drbd0 secondary' terminated with exit code 20
<br>drbdsetup exited with code 20<br><br>I have a truckload of logs here so please tell me what I should provide for further analysis.<br><br>Many thanks!<br><br>Cass Surek<br>