encoding of TSM object names

[gshiman]

As far as I can tell TSM administrative guide doesn't specify encoding of TSM objects and file paths. Neither TSM server option file (dsmsegv.opt) or TSM client options file (tsm.opt) have an encoding option.

Can somebody please explain what encoding is used by TSM server for object names, e.g. management class?
Similarly what encoding is used to record file paths in TSM tape catalog?

Thank you,
George

[moon-buddy]

As far as I can tell TSM administrative guide doesn't specify encoding of TSM objects and file paths. Neither TSM server option file (dsmsegv.opt) or TSM client options file (tsm.opt) have an encoding option.

Can somebody please explain what encoding is used by TSM server for object names, e.g. management class?
Similarly what encoding is used to record file paths in TSM tape catalog?

Thank you,
George

In the absence of any encyption - plain data is recorded.

If by encoding, you mean something like Unicode, then this is the default encoding used by most modern systems.

If you mean encryption, TSM uses AES 128 bit.

[gshiman]

Ed,

I meant character encoding of TSM object names and of file paths.

When you say Unicode, I assume you mean UTF-8, which is character encoding, while Unicode is encoding standard, supporting multiple forms of character encodings, UTF-8 being one of them.

I am aware that UTF-8 character encoding is the default encoding used by many systems.
But I wasn't sure if that is encoding used by TSM server for TSM object names, like management class names.
Can you please confirm if that is indeed the case?

Regarding character encoding of filenames, I knew but forgot in the earlier post that their UTF-8 encoding can be reconstructed from FILE_HEXNAME column in contents table, so I have no question in that regard.

Regards,
George

[moon-buddy]

When I said Unicode, I was really pointing to the standard coding - Unicode that is - that all modern computers adhere to. As to what variant of Unicode IBM uses in TSM, I don't know.

For all we know, TSM might be using proprietory formats which is HEX based. In the original ADSM days, it may well be EBCIDIC in support of the mainframe.

[gshiman]

To continue a bit on on the issue of HEX based filenames in CONTENTS table in TSM tape catalog, I believe that it may represent any encoding of the file name. If given file name on client node was encoded in UTF-8, then HEX-based name will represent UTF-8 multi-byte characters. If the file name was encoded in Latin-1, then HEX-based name will represent Latin-1 multi-byte characters. When reading file names in CONTENTS table there is no way to know what character encoding to use in the file name. UTF-8 should be assumed by default as it offers best chance of success, but there is no guarantee.