AIX dump files can be very big. How to upload them to the IBM ftp-server eCuRep?
When requested to generate diagnostics for AIX operating system analysis the resulting file snap.pax often is very big and causes problems in being uploaded to the IBM ftp-server eCuRep. You can use the 'snapsplit' command to split the snap.pax file into smaller pieces which can be uploaded without problems.
To split the default snap file (snap.pax.Z should be in the current directory), enter the following:
The output of this command is written to current directory.
To split file snap.somefile.pax.Z from system doe, enter the following:
snapsplit -H doe -f snap.somefile.pax.Z
The resulting files will be named snap.machinename.timestamp.pax.Z.xxx where xxx represents the
extension for the split files in the following order : xaa, xab, xac, ..., xba, xbb, ..., xzz
Timestamp is in the format MMDDYYHHMMSS.
The pieces can then be uploaded to the IBM ftp-server eCuRep.
Rename these small files as per eCuRep naming convention to prefix it with the PMR number.
For example: $mv snap.machinename.timestamp.pax.Z.xaa xxxxx.yyy.zzz.snap.machinename.timestamp.pax.Z.xaa
Make sure that you rename all the small files created. The suffix xaa, xab, ... is important to merge the
files to a complete snap.pax afterwards.
You can then upload the files to the IBM ftp-Server eCuRep usinghttps://www.ecurep.ibm.com/app/upload
The files can be merged afterwards by using snapsplit -u , e.g.
to restore a file for which the snap files
(snap.doe.102303141211.xxx) are for system doe, and timestamp 102303141211, type:
snapsplit -u -T 102303141211 -H doe , where doe represents the machinename.
Attention: If any one of the snap files is missing or has been renamed, the snap file created will be corrupted.
snapsplit is located on AIX machines at /usr/sbin/snapsplit