Cisco Prime Infrastructure FTP Backup Issue

[drivera_]

Hi everybody, 

I have the similar issue.  When trying to make a backup with FTP server, I get this message:

CiscoPrimeI/admin# backup BackupFileName repository bakcupCPI
% Creating backup with timestamped filename: BackupFileName-151210-1437.tar.gpg
Backup Started at : 12/10/15 14:37:29
ERROR : Cannot configure the backup directory size settings as the free space available (44975 MB) is less than the current database size(86835 MB).
% Internal error: couldn't create backup file

The FTP server has a 1TB disk free space.  

I created 3 .txt files in the FTP path and when using the command show repository bakcupCPI I get this:

CiscoPrimeI/admin# show repository bakcupCPI
Prueba
1.txt
Prueba
2.txt
Prueba
3.txt
CiscoPrimeI/admin#

It shows that CPI knows where the FTP is located and can access to it, but the issue is still appearing. 

It seems that CPI is viewing the defaultRepo and the ftp repository as a one device:

CiscoPrimeI/admin# dir disk:/defaultRepo

Directory of disk:/defaultRepo

Usage for disk: filesystem
194322432 bytes total used
43146285056 bytes free
45699481600 bytes available
CiscoPrimeI/admin# dir disk:/ftp

Directory of disk:/ftp
4096 Dec 01 2015 23:55:03 reports/
4096 Dec 01 2015 11:19:32 reportsOnDemand/

Usage for disk: filesystem
194322432 bytes total used
43146285056 bytes free
45699481600 bytes available
CiscoPrimeI/admin#

I'll appreciate any help you can give me. 

Diego. 

[Scott Fella]

Make sure that there is enough space on the PI box itself. The backup is created on the PI unit first and then transferred to the FTP server.  Try deleting some old files/logs and make room for the internal backup. 

-Scott

*** Please rate helpful post ***

[tfetzer]

Where would I go in CPI to delete old files & logs?

[ajc · Accepted solution]

By default the configuration files are stored in the following repository so you can delete the ones you do not need. In addition to that the disk can be occupied by crash logs so you need to install root patch and do the following:

repository defaultRepo
  url disk:/defaultRepo
!

prime/admin# show repository defaultRepo
tpi302-16.ubf
prime/admin#

prime/admin# delete disk:/defaultRepo/tpi302-16.ubf

To delete the logs: (you should ask for TAC assistance)

prime/admin# root

Enter root patch password :

Starting root bash shell ...

ade #

ade # du -h --max-depth=6 / opt | grep [0-9  ] G | sort -k2 (TO IDENTIFY AMOUNT OF SPACE LOGS TAKE)

375g /opt

121G /opt/CSCOlumos

99G /opt/CSCOlumos/crash

3.0G /opt/CSCOlumos/da

2.9G /opt/CSCOlumos/da/cdb

------------------------------------------------------------------------

TO list the log files

ade# ls -l /opt/CSCOlumos/crash -t

total 103556676

-rw------- 1 root gadmin 6043954468 dec 10 10:00 java_pid28178.hprof

-rw------- 1 root gadmin 6023298291 dec 1 12:00 java_pid21527.hprof

--------------------------------------------------------------

to remove the logs (SERVICES MUST BE STOPPED BEFORE DOING THIS)

ade# rm -f /opt/CSCOlumos/crash/*

Another option from online documentation is to STOP NCS Application and run the command NCS Cleanup. I do not know the details about this so additional investigation is required.