Sundar Yamunachari wrote: > Moinak Ghosh wrote: > >> The Transfer Module Func. Spec. can be downloaded from >> here: >> >> http://www.opensolaris.org/os/project/livemedia/files/transfermod.pdf >> >> It is currently in the Livemedia files section as I do not have >> access to the Caiman files section. The document should soon >> get moved to the Caiman files page. >> >> Regards, >> Moinak. >> >> _______________________________________________ >> caiman-discuss mailing list >> caiman-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >> >> > Moinak, > > - Section 1 last bullet: > "For the Slim Install initial delivery only a fresh install scenario is > being considered at this time. Upgrade is left as a subsequent activity. > However based on the approach of a LiveCD based installer, upgrades are > typically handled via the packaging system and not directly via the > LiveCD. So for the purposes of the Transfer Module upgrades are out of > scope." > > If slim is planning to use cpio to transfer the contents and not > pkgadd, how will the upgrade work using the packaging system? The > current upgrade works because the information is stored in pkg database > and other places under /var/sadm. The upgrade algorithm uses these > information to decide what needs to be upgraded. > Upgrade is not in the scope for October release. In the future one would upgrade using Snap Upgrade and it will be using the IPS published repository.
> - Section 2 first point: > "Copy the LiveCD contents to disk. This is essentially a straight cpio > of the LiveCD contents from > it's mounted location. The files of course get decompressed as they are > read." > > I assume the cpio archive is created from the contents of the > package. What happends to packaging scripts? pre and post install > scripts and class action scripts? is the pkg database part of the cpio > archive? > Again please note that this transfer module is working with the bits on the Live CD. The bits on the Live CD are a vastly modified version of an installed image. Therefore there will be no pkging information. Furthermore, IPS does not support class action scripts etc. > - Section 2 second point: > "Copy some basic settings from the LiveCD environment to disk. These > include Keyboard selection, Locale selection (on a multi-locale CD/DVD), > Xorg config file if any, the etc/path_to_inst file, the SMF seed > repository, the etc/devices/devid_cache." > > How does the locale stuff works with transfer module? In the current > install, the locale packages corresponding to the list of packages to be > installed will be selected and installed. > > - Section 2 additional tasks: The following may be needed. > > . Do you have any logfiles (or any other files needed for debugging) > created as part of transfer module to be moved to the disk? > . Creating menu.lst file > > Section 5 Interfaces: > > - The api TM_abort_operation() could be changed to TM_abort_transfer() > or TM_cancel_transfer() to clearly indicate that the transfer is > aborted/canceled. > - How do you set the debug? > - Can you expand on the nvlist parameters with the NAME and TYPE of each > parameter? > - Is there any error code returned by TM_perform_transfer() in the case > of error? > - Are you planning to use the same callback structures defined in > orchestrator? Otherwise how can I get the callback information? > > Thanks, > Sundar > > > > > > > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >
