On 08/05/16 09:02 PM, Valeri Galtsev wrote:> > On Sun, May 8, 2016 7:31 pm, Digimer wrote: >> On 08/05/16 08:12 PM, John R Pierce wrote: >>> On 5/6/2016 2:26 PM, Valeri Galtsev wrote: >>>> Which internal hardware RAID controllers will survive some future to >>>> come >>>> in your estimate. First of all my beloved 3ware finally seems to have >>>> passed away. After multiple acquisitions and becoming part of LSI and >>>> getting bought with LSI, it probably became non operational. Namely, >>>> the >>>> latest 3ware cards have ancient firmware. Neither of them supports 4kn >>>> drives. This speaks for itself for me. [Under new ownership] LSI, >>>> though >>>> still having new controllers released, and one of their MegaRAID >>>> controllers (at least) having support for 4kn drives, still may not >>>> last >>>> long (just my feeling, I'd like to hear yours). So, what RAID >>>> controllers >>>> will those of us who like to have hardware RAIDs use in some future to >>>> come? >>> >>> >>> IMHO, "Hardware" (really embedded firmware) RAID is for Windows servers, >>> since MS Windows has awful integrated software raid (aka 'dynamic disk', >>> truly a mess). With Linux, I'd rather use LVM, with BSD, ZFS. >> >> "Hardware RAID" can very well include a controller with dedicated parity >> processing, battery/flash backed write caching and other tangible >> benefits. > > Right, by "hardware RAID" as opposed to a bit more often used term > "software RAID" I did mean the card that has RAID processing done by the > chip on board of the card (parity or in other words modulus 2 sum in case > of RAID-5, and more sophisticated math in case of RAID-6 - I have heard of > at least two algorithms suitable for RAID-6). Thanks, Mr. Digimer, for > clarifying my somewhat vague in this place post.We're not all "Mr".> Any insight, anybody, which hardware RAID cards of rather which > manufacturers of these cards will still make them in a future (say next 5 > years)? Even if you just have feelings, without any thought why, I would > like to hear them. If you prefer to answer off the list, please, e-mail me > directly at galtsev at kicp.uchicago.edu > > Thanks a lot! > > ValeriLSI brand cards are very common across enterprise (I think all tier-1 vendors, except HP, use LSI (now Avago) based controllers. Given that, I would expect their cards will be available for quite some time to come. -- Digimer Papers and Projects: https://alteeve.ca/w/ What if the cure for cancer is trapped in the mind of a person without access to education?
On Sun, May 8, 2016 8:42 pm, Digimer wrote:> On 08/05/16 09:02 PM, Valeri Galtsev wrote: >> >> On Sun, May 8, 2016 7:31 pm, Digimer wrote: >>> On 08/05/16 08:12 PM, John R Pierce wrote: >>>> On 5/6/2016 2:26 PM, Valeri Galtsev wrote: >>>>> Which internal hardware RAID controllers will survive some future to >>>>> come >>>>> in your estimate. First of all my beloved 3ware finally seems to have >>>>> passed away. After multiple acquisitions and becoming part of LSI and >>>>> getting bought with LSI, it probably became non operational. Namely, >>>>> the >>>>> latest 3ware cards have ancient firmware. Neither of them supports >>>>> 4kn >>>>> drives. This speaks for itself for me. [Under new ownership] LSI, >>>>> though >>>>> still having new controllers released, and one of their MegaRAID >>>>> controllers (at least) having support for 4kn drives, still may not >>>>> last >>>>> long (just my feeling, I'd like to hear yours). So, what RAID >>>>> controllers >>>>> will those of us who like to have hardware RAIDs use in some future >>>>> to >>>>> come? >>>> >>>> >>>> IMHO, "Hardware" (really embedded firmware) RAID is for Windows >>>> servers, >>>> since MS Windows has awful integrated software raid (aka 'dynamic >>>> disk', >>>> truly a mess). With Linux, I'd rather use LVM, with BSD, ZFS. >>> >>> "Hardware RAID" can very well include a controller with dedicated >>> parity >>> processing, battery/flash backed write caching and other tangible >>> benefits. >> >> Right, by "hardware RAID" as opposed to a bit more often used term >> "software RAID" I did mean the card that has RAID processing done by the >> chip on board of the card (parity or in other words modulus 2 sum in >> case >> of RAID-5, and more sophisticated math in case of RAID-6 - I have heard >> of >> at least two algorithms suitable for RAID-6). Thanks, Mr. Digimer, for >> clarifying my somewhat vague in this place post. > > We're not all "Mr".Sorry, my usual stupidity... Some time I hopefully learn to be, hm... "wiser"?> >> Any insight, anybody, which hardware RAID cards of rather which >> manufacturers of these cards will still make them in a future (say next >> 5 >> years)? Even if you just have feelings, without any thought why, I would >> like to hear them. If you prefer to answer off the list, please, e-mail >> me >> directly at galtsev at kicp.uchicago.edu >> >> Thanks a lot! >> >> Valeri > > LSI brand cards are very common across enterprise (I think all tier-1 > vendors, except HP, use LSI (now Avago) based controllers. Given that, I > would expect their cards will be available for quite some time to come.Thanks a lot for your insights! This already makes me feel better. In the past LSI would be my definite second choice, and 3ware was winning me only by their transparent web interface. (Several other things LSI had better than 3ware IMHO...) Thanks again! Valeri> > -- > Digimer > Papers and Projects: https://alteeve.ca/w/ > What if the cure for cancer is trapped in the mind of a person without > access to education? > _______________________________________________ > CentOS mailing list > CentOS at centos.org > https://lists.centos.org/mailman/listinfo/centos >++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++
Speaking from experience (I cannot go into detail on this point), and as Digimer pointed out, LSI seems to be the only choice for enterprise level, large scale deployments. If your concern is extremely long term deployments with verifiable data recovery options, software RAID is the only option, as you have strong guarantees that the implementation will never "die" as hardware RAID controllers are likely to do. As others have pointed out, there are indeed tangible benefits to using hardware RAID controllers. It all depends on your use case and project requirements. On May 8, 2016 6:51 PM, "Valeri Galtsev" <galtsev at kicp.uchicago.edu> wrote:> > On Sun, May 8, 2016 8:42 pm, Digimer wrote: > > On 08/05/16 09:02 PM, Valeri Galtsev wrote: > >> > >> On Sun, May 8, 2016 7:31 pm, Digimer wrote: > >>> On 08/05/16 08:12 PM, John R Pierce wrote: > >>>> On 5/6/2016 2:26 PM, Valeri Galtsev wrote: > >>>>> Which internal hardware RAID controllers will survive some future to > >>>>> come > >>>>> in your estimate. First of all my beloved 3ware finally seems to have > >>>>> passed away. After multiple acquisitions and becoming part of LSI and > >>>>> getting bought with LSI, it probably became non operational. Namely, > >>>>> the > >>>>> latest 3ware cards have ancient firmware. Neither of them supports > >>>>> 4kn > >>>>> drives. This speaks for itself for me. [Under new ownership] LSI, > >>>>> though > >>>>> still having new controllers released, and one of their MegaRAID > >>>>> controllers (at least) having support for 4kn drives, still may not > >>>>> last > >>>>> long (just my feeling, I'd like to hear yours). So, what RAID > >>>>> controllers > >>>>> will those of us who like to have hardware RAIDs use in some future > >>>>> to > >>>>> come? > >>>> > >>>> > >>>> IMHO, "Hardware" (really embedded firmware) RAID is for Windows > >>>> servers, > >>>> since MS Windows has awful integrated software raid (aka 'dynamic > >>>> disk', > >>>> truly a mess). With Linux, I'd rather use LVM, with BSD, ZFS. > >>> > >>> "Hardware RAID" can very well include a controller with dedicated > >>> parity > >>> processing, battery/flash backed write caching and other tangible > >>> benefits. > >> > >> Right, by "hardware RAID" as opposed to a bit more often used term > >> "software RAID" I did mean the card that has RAID processing done by the > >> chip on board of the card (parity or in other words modulus 2 sum in > >> case > >> of RAID-5, and more sophisticated math in case of RAID-6 - I have heard > >> of > >> at least two algorithms suitable for RAID-6). Thanks, Mr. Digimer, for > >> clarifying my somewhat vague in this place post. > > > > We're not all "Mr". > > Sorry, my usual stupidity... Some time I hopefully learn to be, hm... > "wiser"? > > > > >> Any insight, anybody, which hardware RAID cards of rather which > >> manufacturers of these cards will still make them in a future (say next > >> 5 > >> years)? Even if you just have feelings, without any thought why, I would > >> like to hear them. If you prefer to answer off the list, please, e-mail > >> me > >> directly at galtsev at kicp.uchicago.edu > >> > >> Thanks a lot! > >> > >> Valeri > > > > LSI brand cards are very common across enterprise (I think all tier-1 > > vendors, except HP, use LSI (now Avago) based controllers. Given that, I > > would expect their cards will be available for quite some time to come. > > Thanks a lot for your insights! This already makes me feel better. In the > past LSI would be my definite second choice, and 3ware was winning me only > by their transparent web interface. (Several other things LSI had better > than 3ware IMHO...) > > Thanks again! > > Valeri > > > > > -- > > Digimer > > Papers and Projects: https://alteeve.ca/w/ > > What if the cure for cancer is trapped in the mind of a person without > > access to education? > > _______________________________________________ > > CentOS mailing list > > CentOS at centos.org > > https://lists.centos.org/mailman/listinfo/centos > > > > > ++++++++++++++++++++++++++++++++++++++++ > Valeri Galtsev > Sr System Administrator > Department of Astronomy and Astrophysics > Kavli Institute for Cosmological Physics > University of Chicago > Phone: 773-702-4247 > ++++++++++++++++++++++++++++++++++++++++ > _______________________________________________ > CentOS mailing list > CentOS at centos.org > https://lists.centos.org/mailman/listinfo/centos >
On 05/08/2016 06:51 PM, Valeri Galtsev wrote:> In the > past LSI would be my definite second choice, and 3ware was winning me only > by their transparent web interface.3ware had a much more pleasant web UI and CLI, but their hardware was terribly unreliable and their performance was pretty awful, too. Today, your options with hardware controllers are mostly LSI with its gawd-awful management software, or Adaptec, or Areca. I rarely see the latter controllers anywhere. They tend to be less expensive than LSI, but they don't benchmark as well, and their smaller market share may fuel doubt about their future prospects. And that, I think, underscores a larger point that people try to make in these conversations, which is: There is no rational case to group hardware RAID controllers together and discuss them exclusively. There are pros and cons to each specific product family and no single quality that disqualifies discussion of other options. That is, the differences between an LSI card and an Adaptec card are no less significant than the differences between an LSI RAID array and a software defined array. My take is this: RAID should not be part of your long-term planning. Everything that's not SAN is moving to software defined storage. Microsoft is moving to Storage Spaces. The UNIX world is moving toward ZFS and btrfs. There are a number of reasons, including hierarchical storage and hybrid storage. Most significant in my opinion though is that while most RAID type can detect spontaneous bit flips, they cannot repair them. You may not use ZFS or btrfs today, but you should definitely be looking at these, long term.
On 05/09/2016 11:01 AM, Valeri Galtsev wrote:> Thanks Gordon! Yes, I know, ZFS, of course. I hear it as you definitely > will use zfs for "bricks" of distributed file system, right?You could, I suppose, but I don't think its use case is limited to that. There aren't many spaces where I think you *shouldn't* plan to use reliable filesystems (ZFS, btrfs, ReFS).
On Mon, May 9, 2016 1:14 pm, Gordon Messmer wrote:> On 05/09/2016 11:01 AM, Valeri Galtsev wrote: >> Thanks Gordon! Yes, I know, ZFS, of course. I hear it as you definitely >> will use zfs for "bricks" of distributed file system, right? > > > You could, I suppose, but I don't think its use case is limited to > that. There aren't many spaces where I think you *shouldn't* plan to > use reliable filesystems (ZFS, btrfs, ReFS).Yes, ZFS, and thanks a lot! For distributed file system "brick" boxes ZFS (btrfs,...) may be a must, but only if distributed filesystem doesn't have its own mechanism ensuring file integrity, right? Valeri ++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++