On Thu, Dec 16, 2010 at 11:36 AM, Matt Gray
I entirely agree with your advocacy in favour of full SSD support for 11.4, i have an SSD in my desktop, but i don't see how it is substantially different from any other rejected feature as a justification for version number changes, if those rejected ones were not good enough already.
Probably because I'm a storage engineer! But also, SSDs are coming down in price. $120 for a decent 40GB SSD right now I believe. It is claimed the Intel Value SSD (40GB ~ $100) will increase desktop performance almost 250%! I think / hope SSD adoption in laptops will be a major transition in 2011. But without discard support, the performance of these devices degrades over time, so a semi-knowledgeable buyer when looking around at Linux Distro's for his new laptop could easily decide to use a distro that claims fully integrated SSD support (including integrated discard support). And since it really does involve a multitude of little package improvements around the distro to get full support, it is something openSUSE can do better than other distros that just keep their packages up to date. Currently, factory (and even 11.3) has 90% of what is needed but it never gets invoked because it requires the user to get in and either manually create cron table entries or edit /etc/fstab. The only new code I see being needed is a Yast enhancement to notice it is partitioning / formatting a SSD and give the user a choice of discard mechanisms. (Or simply force one of the 3 options.) 2 of the 3 are behaviorally similar to defrag and thus require a cron entry (or similar) to invoke them from time to time. I'm assuming Yast could create that cron entry. It would be best if it could done in such a way it happens in the install process of the distro. ie. The installer realizes the target is an SSD and ensures that discard is handled appropriately. Unfortunately, the realtime option (mount --discard) depends on SSD performance that is not yet in production (shipping) drives, so it is almost ignorable at this point, but that may change over the course of 2011. (I've heard rumours there a pre-production drives that can leverage realtime discard for a performance gain.) If it is too late for 11.4 to get those minor yast enhancements, then I personally would like to see that made a goal for 11.5. But it needs to something that gets coordinated support at the top in my opinion. That way the developers, testers, Marketing, PR people are all in the loop. Greg -- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org