I run a script every day to pull out info out of the TSM servers about
backups and one of my co-workers noticed something odd the other day that I did
not see before. The script is just a Unix shell script that among other things
executes a long select statement on each TSM server. Example select:
`dsmadmc -id=tsmop -ps=tsmop -servername=TSM1 -commadelimited SELECT
STATUS.SERVER_NAME,STATUS.SERVER_HLA,NODES.NODE_NAME,NODES.CONTACT,SUMMARY.START_TIME,SUMMARY.EN_TIME,SUMMARY.ACTIVITY,SUMMARY.NUMBER,SUMMARY.ENTITY,SUMMARY.COMMMETH,SUMMARY.ADDRESS,SUMMARY.SCHEDULE_NAME,SUMMARY.EXAMINED,SUMMARY.AFFECTED,SUMMARY.FAILED,SUMMARY.BYTES,SUMMARY.IDLE,SUMMARY.MEDIAW,SUMMARY.PROCESSES,SUMMARY.SUCCESSFUL
FROM STATUS, NODES, SUMMARY WHERE SUMMARY.ENTITY = NODES.NODE_NAME |
clearheaders > $DATADUMP`
Example output:
TSM1.USWASH6,10.82.96.20,I02SS01671-B-IPP,11124279:AP:PWC eGM DWP,2003-11-17
07:38:05.000000,2003-11-17
12:50:15.000000,BACKUP,6154,I02SS01671-B-IPP,Tcp/Ip,10.82.222.24:24256,DAY_STD,993075,5721,79,0,0,0,1,YES
As you can see, this pulls out a bunch of info out of the summary table and
joins it with the nodes table to pull out just the subset of data that I need
regarding backup activity.
The problem I have found is that anytime a backup has any "failed" files, it
puts a "0" in the "bytes transferred field" which is causing many of our
customers to call in asking why we backed up no data for them. I found this
behaviour in both TSM 4.2.4.1 and 5.1.8.1. Should I be pulling this info out
of another table and if so, which one? Why does it put a "0" in the bytes
transferred field when it did transfer data. If I query the act log for a
particular node that failed and look for ANE4961I, I find the corresponding
amount of data transferred but this same value is not populating the summary
table. Hopefully someone can help me out, thanks!
Michael French
Savvis Communications
IDS01 Santa Clara, CA
(408)450-7812 -- desk
(408)239-9913 -- mobile
|