Am Son, 04 Jun 2000 schriebst Du:
>I think I dont understand anythink about drbd ;)
>I dont use heartbeat at this moment ( second step ;) )
>When I've launched the 2 drbd on the servers, all is OK ( with failure, but
>Ok ;):
>
>node 1 and node2 , www and mail have both :
>MASTER_NODE=www:7788
>SLAVE_NODE=mail:7788
>OPTIONS="-r 500"
>SIZE=0
>PROTOCOL=C
>MASTER_IF=192.168.150.3
>SLAVE_IF=192.168.150.4
>MASTER_DEVICE=/dev/nb0
> SLAVE_DEVICE=/dev/nb0
>MASTER_PARTITION=/dev/hdb1
> SLAVE_PARTITION=/dev/hdb1
>MASTER_MNTPOINT=/var/spool/mail
>SLAVE_MNTPOINT=/var/spool/mail
>
>When I launch the both drbd I have on proc/drbd :
>
>www:/etc/rc.d/init.d# cat /proc/drbd
>version : 55
>
>0: cs:Connected st:Secondary ns:0 nr:0 dw:0 dr:0 of:0
>
>on the BOTH node on Secondary state... first strange think.. then I try to
>correct that:
This is ok. DRBD does not know by itself, on which machine you are
going to run the service.
>www:/drbdsetup /dev/nb0 PRI
>
>At this time, I was thinking that drbd was like RAID 1 so, if i change
>somethink on the /var/spool/mail/ on the node1 that will be also changed on
>the node2:
>
Yes, that's right.
( You know http://www.complang.tuwien.ac.at/reisner/drbd/uman.html ?)
But probabely the FS on www is writing the changed blocks a bit later
to the device (e.g. 5 seconds later the directory entry, 30 seconds later
the data blocks, see uman.html!)
>www:/var/spool/mail# touch blah
>
Do you have the Filesystems mounted simultaniously ?
That's not working, because the FS on mail is not rereading the
device, it's using it's cache! Therefore the FS on mail does
not see the blah file.
>mail:/var/spool/mail# ls
>lost+found
>mail:/var/spool/mail#
>
>That dont seems to work..., and the value in /proc/drbd dont change...is it
>normal?
try:
1) connect the device pair. (drbd script)
2) chose one node to run the service (e.g. sendmail)
2a) Switch the device on that node into primary state (on www)
2b) Mount the FS on the device (on www)
2c) Run the service (Modify the content of the FS) (on www)
3) See how mirroring is working
3a) Stop the service (www)
3b) unmount the FS (www)
3c) Switch the device into secondary state (www)
4) Repeate 2a - 2c on the other machine. (on mail)
FYI 2a-2c == datadisk start
3a-3c == datadisk stop
-Philipp
>I think I dont understand anythink about drbd ;)
>I dont use heartbeat at this moment ( second step ;) )
>When I've launched the 2 drbd on the servers, all is OK ( with failure, but
>Ok ;):
>
>node 1 and node2 , www and mail have both :
>MASTER_NODE=www:7788
>SLAVE_NODE=mail:7788
>OPTIONS="-r 500"
>SIZE=0
>PROTOCOL=C
>MASTER_IF=192.168.150.3
>SLAVE_IF=192.168.150.4
>MASTER_DEVICE=/dev/nb0
> SLAVE_DEVICE=/dev/nb0
>MASTER_PARTITION=/dev/hdb1
> SLAVE_PARTITION=/dev/hdb1
>MASTER_MNTPOINT=/var/spool/mail
>SLAVE_MNTPOINT=/var/spool/mail
>
>When I launch the both drbd I have on proc/drbd :
>
>www:/etc/rc.d/init.d# cat /proc/drbd
>version : 55
>
>0: cs:Connected st:Secondary ns:0 nr:0 dw:0 dr:0 of:0
>
>on the BOTH node on Secondary state... first strange think.. then I try to
>correct that:
This is ok. DRBD does not know by itself, on which machine you are
going to run the service.
>www:/drbdsetup /dev/nb0 PRI
>
>At this time, I was thinking that drbd was like RAID 1 so, if i change
>somethink on the /var/spool/mail/ on the node1 that will be also changed on
>the node2:
>
Yes, that's right.
( You know http://www.complang.tuwien.ac.at/reisner/drbd/uman.html ?)
But probabely the FS on www is writing the changed blocks a bit later
to the device (e.g. 5 seconds later the directory entry, 30 seconds later
the data blocks, see uman.html!)
>www:/var/spool/mail# touch blah
>
Do you have the Filesystems mounted simultaniously ?
That's not working, because the FS on mail is not rereading the
device, it's using it's cache! Therefore the FS on mail does
not see the blah file.
>mail:/var/spool/mail# ls
>lost+found
>mail:/var/spool/mail#
>
>That dont seems to work..., and the value in /proc/drbd dont change...is it
>normal?
try:
1) connect the device pair. (drbd script)
2) chose one node to run the service (e.g. sendmail)
2a) Switch the device on that node into primary state (on www)
2b) Mount the FS on the device (on www)
2c) Run the service (Modify the content of the FS) (on www)
3) See how mirroring is working
3a) Stop the service (www)
3b) unmount the FS (www)
3c) Switch the device into secondary state (www)
4) Repeate 2a - 2c on the other machine. (on mail)
FYI 2a-2c == datadisk start
3a-3c == datadisk stop
-Philipp