-What is the best way to test each agent to make sure the installation worked correctly? How to test that the collector is working locally on the machine and possibly test that the TSCO Agent is accessible from the TSCO Gateway Server via the command line? - Unix and Windows Perform Agents -Is there a way to either test remotely or log onto each of the servers and test? -Is the 'best1collect -q -n [nodename]' command the best approach locally on each server |
The 'best1collect' command exists on both the TSCO Gateway Server and the TSCO Agent. So the best1collect command can be executed locally on the machine itself, from one TSCO Agent machine to another TSCO Agent machine, or on the TSCO Gateway Server pointing to a TSCO Agent server. Without a TSCO Gateway Server console there are basically two things that can be tested from the command line:
> $BEST1_HOME/bgs/scripts/best1collect -n [target hostname] -q In these examples, [target hostname] will be the machine 'topten'. Output like this would be good: > $BEST1_HOME/bgs/scripts/best1collect -n topten -q best1collect on topgun: requesting Update Status on topten ... Fri Dec 8 17:16:28 2006 Agent Query Request Starting(/usr/adm/best1_7.3.00/bgs/monitor/log/topten-bgsagent_57300.als) Collect Instance Node Started Started Name Name Name By On ------------ -------------- ---------- ---------- ----------------- ------------ -------------- ---------- ---------- ----------------- *Node: topten has acknowledged Query request successfully. -To start data collection: $BEST1_HOME/bgs/scripts/best1collect -n topten -e+5 -d /tmp -w 1 This will start a 5 minute data collection on the machine with automatic deletion after 60 minutes (1 hour). After 5 or 10 minutes run the following command. $BEST1_HOME/bgs/scripts/best1collect -n topten -Q For example, here is output from a quick test run: > $BEST1_HOME/bgs/scripts/best1collect -n topten -Q best1collect on topgun: requesting Update Status on topten ... Fri Dec 8 17:31:23 2006 Agent Query Request Starting(/usr/adm/best1_7.3.00/bgs/monitor/log/topten-bgsagent_57300.als) Collect Instance Node Started Started Data Length Spill Sample State Config Active Active Term Term LagTime Name Name Name By On Repository HH:MM (Min) (Sec) Groups Data No Data Data No Data (Min) ------------ -------------- ---------- ---------- ----------------- ----------------------------- ------ ----- ------ ------------------ ------ ------ ------- ---- ------- ------- SYSTEM noInstance topten paska Dec-08-2006.17.18 /tmp 0:01 1 10 FILE_AUTO_DELETE 26 14 0 0 12 0 ------------ -------------- ---------- ---------- ----------------- ----------------------------- ------ ----- ------ ------------------ ------ ------ ------- ---- ------- ------- *Node: topten has acknowledged Query request successfully. The output is somewhat hard to read on the command line because it is so long. - There are some things to be aware of when doing a very short collection request like this. If the times are not well synced on the machine that issued the collection request and the machine collecting data the request may never collect or the collection start may be delayed. So, generally, a very short collection should be run on the local machine itself (to ensure there is not a time skew issue). Via the TrueSight Capacity Optimization Gateway Manager useful information regarding the status of data collection can be found under Gateway Manager -> Gateway Reports -> Node History. Using the TSCO Gateway Server version 10.7 and earlier console once computers are included in an active Manager run it becomes much easier to see how well data collection is working in your environment via the UDR Collection Manager (UCM) Status Reports. These reports indicate the success of both data collection and data transfer and will include messages about the source of data collection failures. On the Windows console, they are available under PATROL-Manager => Status Reports. On the Unix console you need to have a browser installed on the console and would open the URL file:///usr/adm/best1_default/local/manager/status/UCMStatus.html (or copy the directory to a another machine with a web browser to view it). The UCM Status Reports are a very useful tool for monitoring (and debugging) data collection and data transfers in your environment. Testing Data TransferIn general, if it is possible to a collection request then data transfer will work since it uses the same communication path. So, it generally isn't necessary to test data transfer separately from data collection.If one wanted to test data transfer from the command line, start by running the following command: $BEST1_HOME/bgs/scripts/best1collect -n [hostname] -p LIST That will list the UDR data available for transfer in the default data repository. If you wanted to look in an alternate location: $BEST1_HOME/bgs/scripts/best1collect -n [hostname] -d /path -p LIST Output will look something like this: $BEST1_HOME/bgs/scripts/best1collect -n abc-xyz-12345 -p LIST best1collect on abc-xyz-12345 : requesting a listing of collected data via the Service Daemon... Node : abc-xyz-12345. Wed Sep 5 11:43:43 2018 VERSION = "10.7.00" ROOT_REPOSITORY_NAME = "/opt/bmc/Patrol3/perform/history" NODE_SELECTOR = "abc-xyz-12345" INSTANCE_SELECTOR = "noInstance" STATUS = 0 STATUS_EXPLANATION = "Success" ---- -------- ---- Node Instance Time ---- -------- ---- abc-xyz-12345 noInstance Sep-04-2018.00.00 ---- -------- ---- *Node: abc-xyz-12345 has acknowledged List request successfully. Then knowing what UDR data is available for transfer (note that UDR data that is still collecting can't be transferred, those are collections that is in UDR_WRITE_ACTIVE state from 'best1collect -Q' output) it can be transferred with the following command: $BEST1_HOME/bgs/scripts/best1collect -n [hostname] -b Mmm-dd-yyyy.hh.mm -p PULL_DELETE For example, from the above output: $BEST1_HOME/bgs/scripts/best1collect -n abc-xyz-12345 -b Sep-04-2018.00.00 -p PULL_DELETE That will transfer the UDR data to the console and will delete it from the remote agent after a successful transfer.
|