EPICURE Software Release Note 14.1
HYDRA's VCS Monitor Program
RDCS node HYDRA is the VAXCluster Console Station for the DISNEY cluster and the front end nodes. It stores the output from the console lines on nodes in the Op Center computer room. Node ARDYH does the same thing for WARNER nodes in the 13th floor computer room. VCS is a package which allows one to easily switch back and forth between nodes, looking at the console output for each. The software lets you look at two nodes at once through windowing, lets you connect to a node so that input goes directly to the chosen node, or lets you view the console output from all of the monitored nodes together in one time stream. The most useful functions will be briefly discussed in this document; for the rest, interested readers are referred to the VAXcluster Console System manual or to the online HELP.
VCS stores the console output data in various files. To view further into the past than is available in the default file, use the following:
REVIEW VCS$ALL_NODES /SINCE=start-date [/BEFORE=end-date]If not specified, the end date will default to the present. You can use the logical VCS$ALL_NODES, as in the example, to get historic data for all of the nodes. If you want data for just one or several nodes, use a comma list of the node names. The REVIEW command merges all of the data into one list, with entries ordered according to the time they were received. Note that there is no way to control-Y or control-C out once you tell VCS to start collecting REVIEW data, so pick a small time window or you may be waiting quite a long time!
or REVIEW MICKEY /SINCE=start-date [/BEFORE=end-date]
Type EXIT or press the F10 key to terminate the VCS program when you are finished, then log out of HYDRA or ARDYH.
VCS stores everything in a disk file. We prefer that you use HYDRA or ARDYH only to look at the console output from a node, and that you don't use VCS to log into the other node at all. VCS is the replacement for the hard-copy paper terminals we used to have; one would read from those but would go out of the way to find a better terminal to login with! Treat this the same way. Otherwise, the system managers will get to read all the commands and all the typos you make, when we read through the log files. We only log in through VCS when we want to make a record of the changes we are making (such as software installations), or when VMS is not running at all on the node (again, software installations or VMS troubleshooting).
The only other good reasons to use VCS to log into a node is when DECNET is down, or VMS is down, or you are doing software installations that will result in VMS going down. In these cases, the console line into a node is the only way to talk to that node, and HYDRA or ARDYH are the only way to reach the console lines.
There are two commands to connect your terminal to one of the monitored nodes. If you use SELECT node-name, the bottom 4 lines of the display will remain, letting you receive notice of other nodes' alarms.
NOTE: For this reason, SELECT is preferred. It reminds you that you are logged in through HYDRA or ARDYH, and reminds you to log out of both your target node as well as HYDRA or ARDYH.
If you use CONNECT node-name, the entire screen is given over to the desired node's console line. This is necessary to run a program which uses the whole screen, but we prefer that you do not run such programs through VCS if there is any other way to connect to the node.
After SELECT and CONNECT, you still have to log into an account on the node in question. A control-G will return you to the VCS monitor program's control. You can CONNECT/SELECT and return with control-G as often as desired; so remember to log out of the remote node when you are through. Then remember to log out of the VCS node too.
Keywords: RDCS, EPICURE, controls, ARDYH, HYDRA, VCS, OPA0:, VAX/VMS, VAXCluster