Hi,
On Tue, Jun 16, 2009 at 02:06:07PM +0200, Sebastian K?sters wrote:
> Hi together,
>
> i am trying to setup a Vmware Server failover installation (on centOS 5) but
> noticed some problems.
>
> My configuration:
>
> Drbd:
>
> ---
>
> resource drbd0 {
> protocol C;
>
> syncer {
> rate 50M;
> }
>
> on kempen.v121.dus1.th {
> device /dev/drbd0;
> disk /dev/sda10;
> address 10.10.121.81:8766;
> meta-disk internal;
> }
>
> on sandiego.v121.fra3.th {
> device /dev/drbd0;
> disk /dev/sda10;
> address 10.10.121.82:8766;
> meta-disk internal;
> }
> }
>
> resource drbd1 {
> protocol C;
>
> syncer {
> rate 50M;
> }
>
> on kempen.v121.dus1.th {
> device /dev/drbd1;
> disk /dev/sda11;
> address 10.10.121.81:8776;
> meta-disk internal;
> }
>
> on sandiego.v121.fra3.th {
> device /dev/drbd1;
> disk /dev/sda11;
> address 10.10.121.82:8776;
> meta-disk internal;
> }
> }
>
> ---
>
> Heartbeat:
>
> haresources:
>
> ---
>
> kempen.v121.dus1.th drbddisk::drbd0 Filesystem::/dev/drbd0::/kempen::ext3
> sandiego.v121.fra3.th drbddisk::drbd1 Filesystem::/dev/drbd1::/sandiego::ext3
>
> ---
>
> Heartbeat and drbd are working without Problems (both newest stable Version).
>
> Then i installed VMware Server on both Servers (also the newest stabel
> Version).
>
> On "kempen.v121.dus1.th" the directory where the vmwaredisks are located is
> configured to /kempen (drbd0)
> On "sandiego.v121.fra3.th" the directory where the vmwaredisks are located is
> configured to /sandiego (drbd1)
>
> The other parameters are default (except user and webports)
>
> I installed windows 2000 in a VM located on "kempen.v121.dus1.th" (i called
> it vmtest.v121.dus1.th)
>
> ---
>
> [r...@kempen /]# ls -al /kempen/vmtest.v121.dus1.th/
> total 1058060
> drwxr-xr-x 4 root root 4096 Jun 16 13:30 .
> drwxrwxrwt 4 root root 4096 Jun 16 10:29 ..
> -rw------- 1 root root 2097152000 Jun 16 13:12
> 564d9bf0-1d3e-a041-36aa-1687643e3c8f.vmem
> drwxrwxrwx 2 root root 4096 Jun 16 13:12
> 564d9bf0-1d3e-a041-36aa-1687643e3c8f.vmem.lck
> -rw------- 1 root root 8684 Jun 16 13:13 vmtest.v121.dus1.th.nvram
> -rw------- 1 root root 917766144 Jun 16 13:30 vmtest.v121.dus1.th.vmdk
> drwxrwxrwx 2 root root 4096 Jun 16 13:30 vmtest.v121.dus1.th.vmdk.lck
> -rw------- 1 root root 0 Jun 16 10:29 vmtest.v121.dus1.th.vmsd
> -rwxr-xr-x 1 root root 2270 Jun 16 13:26 vmtest.v121.dus1.th.vmx
> -rw------- 1 root root 274 Jun 16 10:29 vmtest.v121.dus1.th.vmxf
> -rw-r--r-- 1 root root 52237 Jun 16 13:28 vmware-0.log
> -rw-r--r-- 1 root root 53263 Jun 16 13:26 vmware-1.log
> -rw-r--r-- 1 root root 78665 Jun 16 13:25 vmware-2.log
> -rw-r--r-- 1 root root 52237 Jun 16 13:30 vmware.log
>
> ---
>
> This worked fine.
>
> On "sandiego.v121.fra3.th" i configured a second "Datastore" (over the VMWare
> Webinterface) which i called "kempen" and which points to /kempen
>
> Then i tried to do a failover.
>
> I stopped heartbeat on "kempen.v121.dus1.th". drbd0 failed to
> "sandiego.v121.fra3.th" and is mounted on /kempen
>
> ---
>
> [r...@sandiego /]# df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/sda2 5.9G 890M 4.7G 16% /
> /dev/sda9 3.9G 73M 3.7G 2% /opt
> /dev/sda8 3.9G 73M 3.7G 2% /tmp
> /dev/sda7 4.9G 1.7G 3.0G 35% /usr
> /dev/sda6 9.7G 153M 9.1G 2% /home
> /dev/sda5 15G 298M 14G 3% /var
> /dev/sda1 251M 22M 217M 10% /boot
> tmpfs 16G 0 16G 0% /dev/shm
> /dev/drbd1 37G 177M 35G 1% /sandiego
> /dev/drbd0 37G 1.3G 34G 4% /kempen
>
> ---
>
> The VM files are now located on "sandiego.v121.fra3.th" instead of
> "kempen.v121.fra3.th":
>
> ---
>
> [r...@sandiego /]# ls -al /kempen/vmtest.v121.dus1.th/
> total 1085036
> drwxr-xr-x 4 root root 4096 Jun 16 13:43 .
> drwxrwxrwt 4 root root 4096 Jun 16 10:29 ..
> -rw------- 1 root root 2097152000 Jun 16 13:41
> 564d9bf0-1d3e-a041-36aa-1687643e3c8f.vmem
> drwxrwxrwx 2 root root 4096 Jun 16 13:41
> 564d9bf0-1d3e-a041-36aa-1687643e3c8f.vmem.lck
> -rw------- 1 root root 8684 Jun 16 13:42 vmtest.v121.dus1.th.nvram
> -rw------- 1 root root 917766144 Jun 16 13:43 vmtest.v121.dus1.th.vmdk
> drwxrwxrwx 2 root root 4096 Jun 16 13:41 vmtest.v121.dus1.th.vmdk.lck
> -rw------- 1 root root 0 Jun 16 10:29 vmtest.v121.dus1.th.vmsd
> -rwxr-xr-x 1 root root 2270 Jun 16 13:41 vmtest.v121.dus1.th.vmx
> -rw------- 1 root root 274 Jun 16 10:29 vmtest.v121.dus1.th.vmxf
> -rw-r--r-- 1 root root 52237 Jun 16 13:30 vmware-0.log
> -rw-r--r-- 1 root root 52237 Jun 16 13:28 vmware-1.log
> -rw-r--r-- 1 root root 53263 Jun 16 13:26 vmware-2.log
> -rw-r--r-- 1 root root 78370 Jun 16 13:43 vmware.log
>
> ---
>
> Then i tried to register and start the VM:
>
> ---
> Register:
>
> [r...@sandiego /]# vmrun -T server -h https://localhost/sdk -u root -p test
> register "[kempen] vmtest.v121.dus1.th/vmtest.v121.dus1.th.vmx"
> [r...@sandiego /]#
>
> Start:
>
> [r...@sandiego /]# vmrun -T server -h https://localhost/sdk -u root -p
> test123 start "[kempen] vmtest.v121.dus1.th/vmtest.v121.dus1.th.vmx"
>
> In the webinterface of sandiego VMware asks me if i moved or copied the VM.
>
> I answered: moved / copied (both does not work - why he asks me this?)
>
> On the shell i see:
>
> Error: Unknown error
>
> ---
>
> When in now have a look into "/var/log/vmware/hostd.log"
>
> I get this error:
>
> ---
>
> center: Cannot open the disk
> '/kempen/vmtest.v121.dus1.th/vmtest.v121.dus1.th.vmdk' or one of the snapshot
> disks it depends on.
> Reason: Failed to lock the file.
>
> ---
>
> I then noticed that i have to delete the *.lck files
>
> Is there a better way to do the failover or should i just
> delete all *.lck files after failover and before start?
Probably. But this looks like a question for a vmware forum.
> And is
> there a way to prevend this "did you move oder copy this vm"
> questions?
Of course, whatever the application, it should failover and start
without manual intervention. Again, best to ask vmware people.
Thanks,
Dejan
> Thanks in Advanced!
>
> Mit freundlichen Gr??en / best regards
>
> Sebastian
>
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
