[NTLUG:Discuss] Primary/Primary DRBD ? Anyone

[David Stanaway]

Hey,
is anyone using primary/primary mode in DRBD? I am trying to use that 
for a share ocfs2 filesystem and it works, but I need to manually put 
each node into primary - it doesn't automatically recover after a node 
reboot.

My next project in the same vein was to set up an active/active 
iscsi-target cluster sharing a Primary/Primary DRBD LUN for use with 
multipath iscsi initators. That also will need to automatically recover 
to a primary/primary mode.

One way I could do it is to alter the init scripts to explicitly promote 
the node to primary on boot, but I am not sure how that will play with 
different split brain scenarios.


[node1] debian/etch
/dev/drbd0 = ocfs2 FS cluster for /home with node1/node2 Primary/Primary
/dev/drbd1 = iscsi-target LUN for multipath access from node3 - with 
node1/node2 Primary/Primary

[node2] debian/unstable
/dev/drbd0 = ocfs2 FS cluster for /home with node1/node2 Primary/Primary
/dev/drbd1 = iscsi-target LUN for multipath access from node3 - with 
node1/node2 Primary/Primary

[node3] XP Home
iscsi-initiator multipath with node1/node2 for user data.


I know there are other ways to do this. Consider it a personal proof of 
concept for potential use for other things at work :)


david at yezhov:~$ cat /etc/drbd.conf | sed -e 's/#.*//' -e '/^ *$/d'
global {
     usage-count yes;
}
common {
   syncer { rate 10M; }
}
resource home {
   protocol C;
   handlers {
     pri-on-incon-degr "echo o > /proc/sysrq-trigger ; halt -f";
     pri-lost-after-sb "echo o > /proc/sysrq-trigger ; halt -f";
     local-io-error "echo o > /proc/sysrq-trigger ; halt -f";
     outdate-peer "/usr/sbin/drbd-peer-outdater";
   }
   startup {
     wfc-timeout  12;
     degr-wfc-timeout 120;
   }
   disk {
     on-io-error   detach;
   }
   net {
     allow-two-primaries;
     cram-hmac-alg "sha1";
     shared-secret "boo";
     after-sb-0pri disconnect;
     after-sb-1pri disconnect;
     after-sb-2pri disconnect;
     rr-conflict disconnect;
   }
   syncer {
     rate 10M;
     al-extents 257;
   }
   on yezhov {
     device     /dev/drbd0;
     disk       /dev/YEZHOV-320G/HOME;
     address    10.0.50.6:7788;
     meta-disk  internal;
   }
   on huey {
     device    /dev/drbd0;
     disk      /dev/HUEY-320G/HOME;
     address   10.0.50.5:7788;
     meta-disk internal;
   }
}

david at yezhov:~$ cat /etc/ocfs2/cluster.conf | sed -e 's/#.*//' -e '/^ *$/d'
node:
         ip_port = 7777
         ip_address = 10.0.50.6
         number = 0
         name = yezhov
         cluster = ocfs2
node:
         ip_port = 7777
         ip_address = 10.0.50.5
         number = 1
         name = huey
         cluster = ocfs2
cluster:
         node_count = 2
         name = ocfs2

david at yezhov:~$ cat /etc/ha.d/haresources | sed -e 's/#.*//' -e '/^ *$/d'
yezhov drbddisk::home Filesystem::/dev/drbd0::/home::ocfs2
huey drbddisk::home Filesystem::/dev/drbd0::/home::ocfs2