open a terminal, and type in cvadmin. (If not the root user, you will need to get root access, then issue the command /usr/cvfs/bin/cvadmin)
The cvadmin program will list out what filesystems are currently seen by the metadata servers on that metadata network. They communicate with a service that checks the metadata network for any other servers, and then lists everything it can find (cvfsPM service).
In the cvadmin program, the list will look something like this:
Enter command(s)
For command help, enter "help" or "?".
List FSS
File System Services (* indicates service is in control of FS):
1>*DVS-RT1[0] located on 10.0.0.1
2>*DVS-RT2[0] located on 10.0.0.1
3>DVS-RT1[0] located on 10.0.0.2
4>DVS-RT2[0] located on 10.0.0.2
Basically, the * means that system is the active primary metadata server for that filesystem. In this case, both DVS-RT1 and DVS-RT2 are being controlled by 10.0.0.1 (you might have names instead of IPs... depends if your hosts file was setup that way.)
To fail over the DVS-RT1 filesystem from 10.0.0.1 to 10.0.0.2, while in the cvadmin program, just issue the command "fail 1" (the 1 being the filesystem number, and without the quotes)
It will take a couple of seconds, and you can check by issuing the "select" command. it will re-list what filesystems it sees. Once it fails over, the list will show:
Enter command(s)
For command help, enter "help" or "?".
List FSS
File System Services (* indicates service is in control of FS):
1>*DVS-RT2[0] located on 10.0.0.1
2>*DVS-RT1[0] located on 10.0.0.2
3>DVS-RT2[0] located on 10.0.0.2
4>DVS-RT1[0] located on 10.0.0.1
If you notice, the filsystems have moved position, but DVS-RT1 is now being controlled by 10.0.0.2.
While the manual failover is happening, there could be a couple second interruption of connectivity to your clients, so make sure this happens during a scheduled downtime, or a period of low activity on your SAN.
0 Comments