The perflook script is a tool used to analyze performance behavior of Teradata at the current time. When to use perflook:
How to use Perflook: # /opt/teradata/gsctools/bin/perflook 2018-07-09 13:29:05 - Running /opt/teradata/gsctools/bin/perflook version # 01.00.00.18 2018-07-09 13:29:06 - Running node level collects 2018-07-09 13:29:49 - Running some automated checks for commonly known performance problems Archive and compressing perflook.2018-07-09-1329_1-1 files for downloading... /var/opt/teradata/gsctools/perflook/perflook.2018-07-09-1329_1-1.zip is ready to download. 2018-07-09 13:29:49 - Finished...
The zipped output file is created in the following directory:
If the database appears hung, perflook may also hang. In this case, use
the -d option:
Options in Perflook: -h Print help usage information and exit.
-v Print script version and exit.
-auto Used when perflook is called from another script (ex: canary query), this option is used as a headless option.
-l Enable long verbose mode; stores additional ouput of collected in file puma.TAXlll.
-m Enable memory dumps mode; stores additional output of collected data in files, mm-c and prgtrace.txt.
-d disables collection of output from database utilities, including qrysessn and showlocks
-s Single/Same one node information gathering only Overrides any supplied list of <nodes>
-t Enable stacktraces capture. The output of 'puma -B' is collected. WARNING: Running puma -B can cause a severe impact on system performance.
-B Enable stacktraces capture. The output of 'puma -B' is collected. WARNING: Running puma -B can cause a severe impact on system performance.
nodes An optional comma-separated list of bynet nodes, if not specified data is collected from all nodes on the bynet.
Note: With the -l and/or -m options the output can be huge, these options are not recommended on large sites unless <nodes> list is also provided to narrow the collection scope. <nodes> must be the last argument on the command line. |