Hello Ross, We''re trying to accomplish the same goal over here, ie. serving multiple VMware images from a NFS server. Could you tell what kind of NVRAM device did you end up choosing? We bought a Micromemory PCI card but can''t get a Solaris driver for it... Thanks Gilberto On 7/6/08 9:54 AM, "zfs-discuss-request at opensolaris.org" <zfs-discuss-request at opensolaris.org> wrote:> ------------------------------ > > Message: 6 > Date: Sun, 06 Jul 2008 06:37:40 PDT > From: Ross <myxiplx at hotmail.com> > Subject: [zfs-discuss] Measuring ZFS performance - IOPS and throughput > To: zfs-discuss at opensolaris.org > Message-ID: <21276436.1215351490509.JavaMail.Twebapp at oss-app1> > Content-Type: text/plain; charset=UTF-8 > > Can anybody tell me how to measure the raw performance of a new system I''m > putting together? I''d like to know what it''s capable of in terms of IOPS and > raw throughput to the disks. > > I''ve seen Richard''s raidoptimiser program, but I''ve only seen results for > random read iops performance, and I''m particularly interested in write > performance. That''s because the live server will be fitted with 512MB of > nvram for the ZIL, and I''d like to see what effect that actually has. > > The disk system will be serving NFS to VMware to act as the datastore for a > number of virtual machines. I plan to benchmark the individual machines to > see what kind of load they put on the server, but I need the raw figures from > the disk to get an idea of how many machines I can serve before I need to > start thinking bigger. > > I''d also like to know if there''s any easy way to see the current performance > of the system once it''s in use? I know VMware has performance monitoring > built into the console, but I''d prefer to take figures directly off the > storage server if possible. > > thanks, > > Ross >
On Sun, Jul 06, 2008 at 08:36:33PM -0400, Gilberto Mautner wrote:> We''re trying to accomplish the same goal over here, ie. serving > multiple VMware images from a NFS server. > > Could you tell what kind of NVRAM device did you end up choosing? We > bought a Micromemory PCI card but can''t get a Solaris driver for it...A Solaris driver definitely exists for the uMem PCI cards. Don''t know how you''d get it though but some performance stats were done with one of the their PCI cards on Solaris: http://blogs.sun.com/perrin/entry/slog_blog_or_blogging_on> Gilberto > > > On 7/6/08 9:54 AM, "zfs-discuss-request at opensolaris.org" > <zfs-discuss-request at opensolaris.org> wrote: > > > ------------------------------ > > > > Message: 6 > > Date: Sun, 06 Jul 2008 06:37:40 PDT > > From: Ross <myxiplx at hotmail.com> > > Subject: [zfs-discuss] Measuring ZFS performance - IOPS and throughput > > To: zfs-discuss at opensolaris.org > > Message-ID: <21276436.1215351490509.JavaMail.Twebapp at oss-app1> > > Content-Type: text/plain; charset=UTF-8 > > > > Can anybody tell me how to measure the raw performance of a new system I''m > > putting together? I''d like to know what it''s capable of in terms of IOPS and > > raw throughput to the disks. > > > > I''ve seen Richard''s raidoptimiser program, but I''ve only seen results for > > random read iops performance, and I''m particularly interested in write > > performance. That''s because the live server will be fitted with 512MB of > > nvram for the ZIL, and I''d like to see what effect that actually has. > > > > The disk system will be serving NFS to VMware to act as the datastore for a > > number of virtual machines. I plan to benchmark the individual machines to > > see what kind of load they put on the server, but I need the raw figures from > > the disk to get an idea of how many machines I can serve before I need to > > start thinking bigger. > > > > I''d also like to know if there''s any easy way to see the current performance > > of the system once it''s in use? I know VMware has performance monitoring > > built into the console, but I''d prefer to take figures directly off the > > storage server if possible. > > > > thanks, > > > > Ross > > > > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >-- albert chin (china at thewrittenword.com)
Hi Gilberto, I bought a Micro Memory card too, so I''m very likely going to end up in the same boat. I saw Neil Perrin''s blog about the MM-5425 card, found that Vmetro don''t seem to want to sell them, but then then last week spotted five of those cards on e-bay so snapped them up. I''m still waiting for the hardware for this server, but regarding the drivers, if these cards don''t work out of the box I was planning to pester Neil Perrin and see if he still has some drivers for them :) The cards were only ?20 each, I figured it was a bit of a gamble buying them, but hopefully one that will pay off. Ross This message posted from opensolaris.org
Ross wrote:> Hi Gilberto, > > I bought a Micro Memory card too, so I''m very likely going to end up in the same boat. > I saw Neil Perrin''s blog about the MM-5425 card, found that Vmetro don''t seem to want > to sell them, but then then last week spotted five of those cards on e-bay so snapped > them up. > > I''m still waiting for the hardware for this server, but regarding the drivers, if these > cards don''t work out of the box I was planning to pester Neil Perrin and see if he still > has some drivers for them :)Unfortunately, there are a couple of problems: 1. It''s been a while since I used that board and driver. I recently tried pkgadd-ing on the latest Nevada build and it hung. I''m not sure if the latest Nevada is somehow incompatible. I didn''t have time to track down the cause. 2. I received the board and driver from another group within Sun. It would be better to contact Micro Memory (or whoever took them over) directly, as it''s not my place to give out 3rd party drivers or provide support for them. Sorry for the bad news: Neil.
Hi Neil, No problem, thanks for the info. I knew these cards were a gamble, but if I can get them working, it will be worth it. I''ve today spoken to the UK office of VMetro about drivers but I''m not holding out too much hope. They were very friendly and polite, but explained that they simply don''t support these for end users, it''s for large OEM''s only. However, I did do a bit of digging this morning and found the e-mail address of Micro Memory''s lead software developer, who looks to be the chap responsible for developing the Solaris drivers in the first place. So I''ll be dropping him a line shortly and seeing if he can help at all. Ross This message posted from opensolaris.org
>>>>> "r" == Ross <myxiplx at hotmail.com> writes: >>>>> "np" == Neil Perrin <Neil.Perrin at Sun.COM> writes:np> 2. I received the board and driver from another group within np> Sun. It would be better to contact Micro Memory (or whoever np> took them over) directly, as it''s not my place to give out 3rd np> party drivers or provide support for them. Then hopefully when Sun releases their new batch of SSD devices, they will release source for the full driver stack under a redistributable license so that no well-meaning geek has to be in your awkwardly unhelpful position, caught between obligations of NDA/copyright/``place'''' and the basic and reasonable obligations necessary to maintain a ``community''''. I''ve heard Sun people at users'' groups promise that all new Solaris subsystems will include source, but so far this doesn''t apply to hardware, not even to the hardware Sun sells. In this case source would solve (1) and (2) because you''d be (2) free to redistribute whatever you had a month ago, and Ross would (1) have a fighting chance of forward-porting the driver he got from you. This isn''t the case for existing Sun disk drivers that I know about like the X4500 SATA chip or the LSI Logic mpt RAID card in SPARC SATA systems, while Linux and I think BSD have free software drivers for both chips---at best the Sun drivers are (2) redistributable, and I''m not even clear on that because it''s surprisingly tricky to determine. r> they simply don''t support these for end users, it''s for large r> OEM''s only. [...] found the e-mail address of Micro Memory''s r> lead software developer, who, unlike the salespeople, will probably understand the obvious difference between providing ``support,'''' and taking the basic responsibility to either archive all downloadables that aren''t redistributable, or make them redistributable if they don''t want to track them any more, but who probably won''t be in a position to help you any more than Neil is. If their contractor did give you the drivers, would you avoid mentioning it here for fear a bunch of other people would ask you for copies, putting you in the same awkward spot? Would you justify the reticence by thinking you were hiding the drivers from us out of loyalty and ``gratitude'''' to the contractor who wrote them? It stinks, and I recognize the smell. We''ve been here before. I ought to have better things to do with my life than pirating software to support obscure proprietary abandonware (but apparently not better than writing emails whining about the situation). -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 304 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20080709/f4321d64/attachment.bin>
I think the problem Miles is that this isn''t Sun hardware, and I completely understand that as a Sun employee, Neil really can''t be seen to distribute something that''s untested and unsupported, and quite possibly under NDA. On the other hand, if I get hold of these drivers, I''m under no such obligation and I''ll be happily making them available for everybody who wants them. I already know of two other people who are keen to get these and I''m sure there are others. These cards are starting to show up on the second hand market now, finding a set of Solaris drivers would be a welcome bonus for a good few people. This message posted from opensolaris.org
On Wed, Jul 9, 2008 at 1:14 PM, Ross <myxiplx at hotmail.com> wrote:> I think the problem Miles is that this isn''t Sun hardware, and I completely > understand that as a Sun employee, Neil really can''t be seen to distribute > something that''s untested and unsupported, and quite possibly under NDA. > > On the other hand, if I get hold of these drivers, I''m under no such > obligation and I''ll be happily making them available for everybody who wants > them. I already know of two other people who are keen to get these and I''m > sure there are others. > > These cards are starting to show up on the second hand market now, finding > a set of Solaris drivers would be a welcome bonus for a good few people. > > > This message posted from opensolaris.org > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >Do we have drivers available for ANY OS for these cards currently? It''d be nice to at least be able to test if they function properly. --Tim -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20080709/e28e127b/attachment.html>
>>>>> "r" == Ross <myxiplx at hotmail.com> writes:r> I think the problem Miles is that this isn''t Sun hardware In this case it''s not, but please do not muddle my point: Marvell SATA and LSI Logic mpt SATARAID and many other (most?) drivers have the same problem. Right now there are, AIUI: * closed-source non-redistributable drivers (SXCE only) * closed-source redistributable drivers (SXCE, Indiana, Nexenta) * open-source redistributable drivers (SXCE, Indiana, Nexenta) The logical fourth category of open-source non-redistributable drivers doesn''t exist---you CAN have a non-redistributable, $0 driver which includes source code, but it wouldn''t meet the open source specification. The word ``third-party driver'''' is thrown around a lot. I guess it was a common word in the pre-Opensolaris days? The three categories are orthogonal to bundling or support entitlements, and there are plenty of Solaris/SXCE-bundled, support-entitled drivers in the first category. r> I completely understand that as a Sun employee, Neil really r> can''t be seen to distribute something that''s untested and r> unsupported, and quite possibly under NDA. AIUI it''s not personal, or be-seen-as. It''s, do you have the right to do it, or do you not. For example, I do not have the right to give you an SXCE DVD. You have to download it yourself. (hope you don''t need an old version!) I DO have the right to give you a Nexenta or OpenSolaris 2008.05 DVD. This is redistribution. To pass redistribution rights on to me, Sun left drivers out of the OpenSolaris/Indiana release and Nexenta out of the Nexenta release. And just as Micro Memory can take a formerly-$0 driver down from their web page, Sun can take down the SXCE b12345 .iso, and if you don''t already have a copy hoarded you''re not technically allowed to have your friend copy his DVD and give it to you. r> if I get hold of these drivers, I''m under no such obligation The obligation would come when you get the drivers---you''ll be given drivers on the condition you agree to something. Since you don''t have them yet, you''re in a bad position to promise this. You could promise, ``I won''t accept drivers from anyone who makes me promise not to redistribute them or not to release the source code of them,'''' (or publish benchmarks without the manufacturers approval COUGH COUGH) which is what I _wish_ Sun would do to the chip and card vendors from which they get components in the hardware they sell, but they don''t. You could also promise, ``If someone makes me agree not to redistribute this, I''ll agree and then break the agreement, because I care more about preserving the community than I do about respecting legal agreements.'''' To me it seems like technical people take exclusively the former approach, and casual non-technical users almost exclusively the latter. I guess there are a lot of people in the world who can repeatedly make the latter statement publicly without hurting their careers, but maybe not many such people on this list. anyway sorry it''s OT. I''ll drop it now. I should hunt for a license-discuss at opensolaris.org list or something. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 304 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20080709/ab455cbd/attachment.bin>