Исходное сообщение
"Доступна финальная версия ZFS для Linux, работающая на уровн..." Отправлено Аноним, 28-Янв-11 14:56
KQ Stor™ Known defects, limitations and work arounds with the core ZFS operations and features. 1. block and char devices Accessing block and character devices over zfs filesystem are not supported. 2. zfs rollback panic Under certain conditions zfs rollback might panic the system. Please backup your data before attempting this and we recommend that you reboot the machine and run zpool scrub after the rollback. 3. Migrating data between Solaris and Linux Although the pool layout in this port is exactly the same as that on Opensolaris, there are subtle differences. Firstly ACLs are not supported and fuids are never populated in Linux. Please contact us before you attempt to import a filesystem from Solaris to Linux. Secondly the extended attributes in Linux are implemented by prefixing the attribute name with the attribute class. e.g. "user." for user attributes "security." for SELinux security attributes. This imposes the restrition that any arbitrary named attribute will not exist on Linux, it will have a prefix either "user" or "security". While the attributes in Opensolaris can have any arbitrary name. NFS shares related Know problems and work arounds related export ZFS filesystems over NFS. 1. cannot access .zfs/snapshots over nfs This limitation will be addressed in future releases. The snapshots directories can only be accessed from the host machine or if the snapshot mount point has been explicitly exported. 2. zfs share does not work zfs share command does not export the filesystem marked with sharenfs=on. The exporting is done at the time zfs set sharenfs=on dataset is executed. Since this internally invokes the exportfs command the effects are equivalent, if you chose to use exportfs instead of zfs sharenfs=on to export. This current limitation of "zfs share" not invoking exportfs will be fixed in a minor patch release. 3. How do I set nfs options while exporting Currently the nfs integration within the zfs command is limited to invoking exportfs. It does not give control to set the various options like "no_root_squash" or other exportfs options. If such fine grained control over the nfs export is required we recommend that you use exportfs directly. 4. Filesystem exported via /etc/fstab doesn't show correctly on client Depending on your setup the nfs exports might happen before the zfs filesystems are mounted. This would result in mountpoints for the ZFS filesystems being exported instead. Unexport and re-export the /etc/fstab entries. Release Notes 2 Other Linux features 1. io_submit(2) libaio.h Posix AIO is fully supported. In the current release Linux AIO support has been disabled. It should be enabled in one of the minor releases

Ваше сообщение
Имя*:
EMail:	Для отправки ответов на email укажите знак ! перед адресом, например, !user@host.ru (!! - не показывать email). Более тонкая настройка отправки ответов производится в профиле зарегистрированного участника форума.
Заголовок*:
Сообщение*:	> KQ Stor™ > Known defects, limitations and work arounds with the core ZFS operations and > features. > 1. block and char devices > Accessing block and character devices over zfs filesystem are not supported. > 2. zfs rollback panic > Under certain conditions zfs rollback might panic the system. Please backup your > data before > attempting this and we recommend that you reboot the machine and run > zpool scrub after the > rollback. > 3. Migrating data between Solaris and Linux > Although the pool layout in this port is exactly the same as > that on Opensolaris, there are subtle > differences. Firstly ACLs are not supported and fuids are never populated in > Linux. Please contact > us before you attempt to import a filesystem from Solaris to Linux. > Secondly the extended attributes > in Linux are implemented by prefixing the attribute name with the attribute > class. e.g. "user." > for user attributes "security." for SELinux security attributes. This imposes the restrition > that any > arbitrary named attribute will not exist on Linux, it will have a > prefix either "user" or "security". > While the attributes in Opensolaris can have any arbitrary name. > NFS shares related > Know problems and work arounds related export ZFS filesystems over NFS. > 1. cannot access .zfs/snapshots over nfs > This limitation will be addressed in future releases. The snapshots directories can > only be accessed > from the host machine or if the snapshot mount point has been > explicitly exported. > 2. zfs share does not work > zfs share command does not export the filesystem marked with sharenfs=on. The > exporting is > done at the time zfs set sharenfs=on dataset is executed. Since this > internally invokes the exportfs > command the effects are equivalent, if you chose to use exportfs instead > of zfs sharenfs=on to > export. This current limitation of "zfs share" not invoking exportfs will be > fixed in a minor patch > release. > 3. How do I set nfs options while exporting > Currently the nfs integration within the zfs command is limited to invoking > exportfs. It does not > give control to set the various options like "no_root_squash" or other exportfs > options. If such fine > grained control over the nfs export is required we recommend that you > use exportfs directly. > 4. Filesystem exported via /etc/fstab doesn't show correctly on client > Depending on your setup the nfs exports might happen before the zfs > filesystems are mounted. > This would result in mountpoints for the ZFS filesystems being exported instead. > Unexport and > re-export the /etc/fstab entries. > Release Notes > 2 > Other Linux features > 1. io_submit(2) libaio.h > Posix AIO is fully supported. In the current release Linux AIO support > has been disabled. It should > be enabled in one of the minor releases
	Введите код, изображенный на картинке: