jan damborsky wrote: > Hi all, > > could I please ask for reviewing the fix for following CR (blocker for > 2008.11) ? > > 1771 Installer can't be restarted if it already created ZFS root pool > > * bug > http://defect.opensolaris.org/bz/show_bug.cgi?id=1771 > > * Webrev: > http://cr.opensolaris.org/~dambi/bug-1771/ > > > Thank you very much, > Jan > > > The modules affected and tested: > * libti, liborchestrator, libict > > test conditions: > * platform: > - vmWare (800 MiB RWM for guest OS) on W2110Z (2GiB RWM) > - Ultra 20 (1GiB RWM) > > * ISO: OpenSolaris based on build 98 created using Distro > Constructor taking advantage of checkpointing feature > > tests carried out: > > [1] Regression tests > * fresh installation on minimum disk size (without swap) > * fresh installation on minimum disk size (with swap) > * fresh installation on recommended disk size > > [2] Test of new functionality > * restart of interrupted fresh installation on recommended disk size > - swap and dump created > - process interrupted during transfer phase > - when restarted, installer successfully released > dump and swap ZFS volumes, ZFS pool was destroyed, > installation finished successfully > > * restart of finished fresh installation on recommended disk size > - swap and dump created > - installation process finished successfully > - when restarted after successful installation, installer > refused to destroy the ZFS pool > > * attempt to run installer, when 'rpool' exists, but not marked 'busy' > - ZFS pool 'rpool' manually created (not marked 'busy') > - installer refused to destroy existing ZFS pool > > > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
Jan, This looks good. I see two possible issues described below. Joe usr/src/lib/libti/test_ti.c 43 lines changed: +-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-= Allowing both upper and lower case P for target_type pool create/release could be confusing. Perhaps a clearer description in the usage would help. usr/src/lib/libti/ti_api.h 15 lines changed: +-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-= 87 #define RPOOL_PROPERTY_STATE "org.opensolaris.caiman:install" 88 #define RPOOL_BUSY "busy" 89 #define RPOOL_READY "ready" 90 91 /* 92 * ZFS volume names for swap and dump 93 */ 94 #define ZFS_VOL_NAME_SWAP "swap" 95 #define ZFS_VOL_NAME_DUMP "dump" Shouldn't these have the TI_ prefix?
