Monitoring and Cleaning Disk Space For Cognos Consumer Insight Data Node

Answer

How to Monitor the Data Node

1. Log into the data node (for instance using putty) using user cciusr

2. Check for available disk system space by executing
df –k

3. Look for Use% of the local file system

4. If this reaches or exceeds 90%, perform the cleanup procedure that follows.

1. As user cciusr execute
cd /local/cci/prod/dls/data/staging

2. The directory contains a list of tenants

3. For each tenant directory under /local/cci/prod/dls/data/staging

a. Check the contents of the adhoc subdirectory

b. If the directory adhoc contains only one subdirectory no action is required and goto Step 4.

c. If the directory adhoc contains more than one subdirectory, goto the adhoc subdirectory
cd <tenant name>/adhoc

d. Type
ll –tr

e. Delete all directories except the newest one. The newest one is the one at the end of the list
rm -rf <old directory>


4. For each tenant directory under /local/cci/prod/dls/data/staging

a. Check the contents of the scheduled subdirectory

b. If scheduled contains only one directory you have completed this procedure

c. If scheduled contains more than one subdirectory goto scheduled directory
cd scheduled

d. Identify old directories by executing
ll –tr

e. Delete all directories except the newest one. The newest one is the one at the end of the list
rm -rf <old directory>

1. As user cciusr execute goto the following directory
cd /local/cci/prod/reports
if reports folder exists, move to step 2 below

2. The directory contains exported csv files and corresponding sql files for each tenant, execute
du . -h -c |grep total

3. Goto each tenant directory, and execute
rm -rf *

4. Clean up dataload log files:
cd /local/cci/prod/dls/services/Dataloader

5. List logs files
ls -lt dataloader.log*

6. Remove log files older than 10 days:
find . -name "dataloader.log*" -mtime +10 -exec rm {} \;