Hi Angela,
angela wrote: > Thanks Jan for detailed review. > You are welcome :-) > According to your feedback, there are two aspects need to do: > 1. add sparc support > 2. modify zfs pool creation on slice device > > I would do aspect 2 firstly and then put back the testsuite into stc, > and then add sparc support. > If you would agree, I might recommend to focus on support for Sparc first, since this is higher priority at time of being. We have been covering/testing x86 platform for a while, but as far as Sparc is concerned we stay at the beginning. As we did initial testing for Sparc, we don't expect to find serious issues modulo those already reported. However, more testing would need to be done in order to be sure we didn't neglect anything important. Thank you, Jan > Thanks, > Angela > > jan damborsky ??: > >> Hi Angela, >> >> I have gone through the list of test assertions (attached for reference). >> They seem to cover the latest set of Target Instantiation features >> pretty well. I have only couple comments - please see below. >> >> I am also including William, who worked on libti support for >> Distro Constructor. >> >> As far as set of supported platforms is concerned, we have >> been focusing on x86 so far. Since we are now in process >> of enhancing the scope by taking Sparc platform into account, >> could I please ask you, if it might be possible to extend testing >> procedures, so that Sparc platform could be covered as well ? >> >> In general, libti is supposed to work across all Sparc architectures, >> but in near future 'sun4v' is prefered, so if you would like to limit >> the scope, please focus on 'sun4v' at time of being. >> >> As fdisk related test assertions are not applicable for Sparc, >> following test cases would be omitted: >> >> create_fdisk_001 >> create_fdisk_002 >> create_fdisk_003 >> >> With respect to the issues affecting support for Sparc we are currently >> aware of following problem: >> >> 3136 TI fails to create VTOC on unlabeled disk on Sparc >> >> If the unlabeled disk is attached to the SUT, it is necessary >> to create label on it (using format(1M)), otherwise libti would >> fail to create VTOC on that drive. >> >> Thank you very much, >> Jan >> >> >> create_vtoc_002 >> --------------- >> Since only SMI labeling is currently supported and utilized >> in the installer, this test case is assumed to work, but >> potential issues seen there would not be considered as libti >> bugs at time of being. >> >> >> create_rpool_004 >> create_rpool_001n >> create_be_001 >> create_be_001n >> create_fs_001 >> ----------------- >> Since creating ZFS pool on whole disk is unsupported >> and not currently utilized in the installer, could >> you please modify these test assertion, so that ZFS >> pool is created on slice instead ? >> >> >> create_volume_001 >> ----------------- >> # - ZFS root pool's name, $ZFS_RPOOL_NONEXIST, other than the >> # existing pool(s), check with 'zpool status'. >> -> >> # - name of existing ZFS root pool, $ZFS_RPOOL_NAME, check with >> # 'zpool status'. >> >> >> # - First, '$DRIVER_PATH/$DRIVER_NAME -c -t p -p $ZFS_RPOOL_NONEXIST' >> # - Second, '$DRIVER_PATH/$DRIVER_NAME -c -t l -p $ZFS_RPOOL_NONEXIST >> # -n $ZFS_VOLUME_NAME -z $ZFS_VOLUME_SIZE -u $ZFS_VOLUME_TYPE' >> -> >> # - First, '$DRIVER_PATH/$DRIVER_NAME -c -t p -p $ZFS_RPOOL_NAME >> # -d $SLICE_NAME' >> # - Second, '$DRIVER_PATH/$DRIVER_NAME -c -t l -p $ZFS_RPOOL_NAME >> # -n $ZFS_VOLUME_NAME -z $ZFS_VOLUME_SIZE -u $ZFS_VOLUME_TYPE' >> >> >> create_fs_001 >> ------------- >> # - ZFS root pool's name, $ZFS_RPOOL_NONEXIST, other than the >> # existing pool(s), check with 'zpool status'. >> -> >> # - name of existing ZFS root pool, $ZFS_RPOOL_NAME, check with >> # 'zpool status'. >> >> >> # - First, '$DRIVER_PATH/$DRIVER_NAME -c -t p -d $DISK_NAME >> # -p $ZFS_RPOOL_NONEXIST' >> # - Second, '$DRIVER_PATH/$DRIVER_NAME -c -t m -p $ZFS_RPOOL_NONEXIST >> # -n $ZFS_FS_NAME' >> -> >> # - First, '$DRIVER_PATH/$DRIVER_NAME -c -t p -d $SLICE_NAME >> # -p $ZFS_RPOOL_NAME' >> # - Second, '$DRIVER_PATH/$DRIVER_NAME -c -t m -p $ZFS_RPOOL_NAME >> # -n $ZFS_FS_NAME' >> >> >> release_zfspool_001 >> ------------------- >> # - '$DRIVER_PATH/$DRIVER_NAME -c -t p -p $ZFS_RPOOL_NAME -R' >> -> >> # - '$DRIVER_PATH/$DRIVER_NAME -c -t P -p $ZFS_RPOOL_NAME' >> >> >> release_zfspool_001n >> -------------------- >> # - '$DRIVER_PATH/$DRIVER_NAME -c -t p -p $ZFS_RPOOL_NONEXIST -R' >> -> >> # - '$DRIVER_PATH/$DRIVER_NAME -c -t P -p $ZFS_RPOOL_NONEXIST' >> >> >> angela wrote: >> >> >>> Thanks, Jan. >>> >>> I attached the latest test assertion, please review. >>> >>> And also the workspace is at: >>> /net/rainday.prc/export/Projects/indiana/General/ws-snv-itc/src/suites/libti >>> >>> Cheers, >>> Angela >>> >>> jan damborsky ??: >>> >>> >>>> Hi Angela, >>>> >>>> only couple of things changed since last release - >>>> I think (I am not sure) that Hao did some changes >>>> in the testsuite to accommodate it. >>>> If you agree, I might take a look at latest list of test >>>> cases and I would get back to you with comments >>>> what should be changed or enhanced. >>>> Please let me know, what you think about this. >>>> >>>> Thank you, >>>> Jan >>>> >>>> >>>> angela wrote: >>>> >>>> >>>> >>>>> Hi Jan, >>>>> >>>>> This is Angela from Install Beijing QE team, I'm responsible for libti >>>>> testsuite now, replacing Wang Hao who has transferred to BSTE team. >>>>> >>>>> I heard that you wrote libti test driver, so I wonder does there any >>>>> modification on it, since last release? >>>>> And does there any new features on libti? If yes, I will enhance >>>>> libti testsuite accordingly. >>>>> >>>>> Thanks, >>>>> Angela >>>>> >>>>> >>>>> >>>>> >> >> > >
