December 17, 201510 yr I wonder if dual parity will slow down writes to the array significantly. Cannot see why it should as one can be writing (or reading) 2 drives in parallel. However I guess until we actually have 6.2 to try we will not know for certain.
December 17, 201510 yr I wonder if dual parity will slow down writes to the array significantly. While it will increase the number of disk I/O's required for a write, I wouldn't think it will cause a significant drop in write speed, as the additional I/O's can be done in parallel. i.e. today it needs to do 2 reads (the parity drive and the drive being written to), followed by 2 writes (to the same 2 drives). With dual parity that will be 3 and 3. There key difference is that it could take a bit longer to do the initial 3 reads, since the drives aren't likely to all be in rotational sync (thus a bit more rotational latency). Of course if you're writing to a cache, it won't matter at all, since the actual writes to the array will be done later when the mover runs. Note, by the way, that the testing shown in the video is ALL done with cache writes ... the actual array has no assigned parity and a single array drive. All of the other drives are assigned to the cache. Also note that although he SAYS he has a bunch of enterprise 8TB drives to use for this, they are actually 6TB drives.
December 17, 201510 yr I wonder if dual parity will slow down writes to the array significantly. While it will increase the number of disk I/O's required for a write, I wouldn't think it will cause a significant drop in write speed, as the additional I/O's can be done in parallel. i.e. today it needs to do 2 reads (the parity drive and the drive being written to), followed by 2 writes (to the same 2 drives). With dual parity that will be 3 and 3. There key difference is that it could take a bit longer to do the initial 3 reads, since the drives aren't likely to all be in rotational sync (thus a bit more rotational latency). Of course if you're writing to a cache, it won't matter at all, since the actual writes to the array will be done later when the mover runs. Note, by the way, that the testing shown in the video is ALL done with cache writes ... the actual array has no assigned parity and a single array drive. All of the other drives are assigned to the cache. Also note that although he SAYS he has a bunch of enterprise 8TB drives to use for this, they are actually 6TB drives. If you watch the video, Linus is clear the 8TB are used in two other NAS to allow him to clear the 45 bay system for testing. The 8TB are not used for testing.
December 28, 201510 yr And here is the latest video where you get some decent screen shots of 6.2 beta with dual parity support...
December 29, 201510 yr And here is the latest video where you get some decent screen shots of 6.2 beta with dual parity support... So the gui is available on the local monitor with 6.2? That login screen looked local to me, and he talks about using the gui locally, and not just on the vm.
December 29, 201510 yr So the gui is available on the local monitor with 6.2? That login screen looked local to me, and he talks about using the gui locally, and not just on the vm. Nothing confirmed, but definitely looks like it. The way Linus was talking, needing the screen local to setup, looks like we are getting a local GUI interface.
December 29, 201510 yr So the gui is available on the local monitor with 6.2? That login screen looked local to me, and he talks about using the gui locally, and not just on the vm. Nothing confirmed, but definitely looks like it. The way Linus was talking, needing the screen local to setup, looks like we are getting a local GUI interface. I don't see anything resembling a local GUI to setup unraid 6.2 in that video. I see Asus's UEFI BIOS. I see a Windows 10 OS running Chrome browser to the other unRAID system, likely from one of the 2 laptops also on the same table. The 2 other laptops can be seen at the 12:21 mark of the video. Later in the video he is using video card passthrough to display the video images from the virtual machine to the monitor.
December 29, 201510 yr So the gui is available on the local monitor with 6.2? That login screen looked local to me, and he talks about using the gui locally, and not just on the vm. Nothing confirmed, but definitely looks like it. The way Linus was talking, needing the screen local to setup, looks like we are getting a local GUI interface. I don't see anything resembling a local GUI to setup unraid 6.2 in that video. I see Asus's UEFI BIOS. I see a Windows 10 OS running Chrome browser to the other unRAID system, likely from one of the 2 laptops also on the same table. The 2 other laptops can be seen at the 12:21 mark of the video. Later in the video he is using video card passthrough to display the video images from the virtual machine to the monitor. I suspect if there is a local gui, linus was told not to demo it. However, there are several hints, including the unraid login screen that looks like a gui login at 6:33, and again later. I suppose that could just be the new password webui screen that linus had full screen, but I doubt it.
December 29, 201510 yr He talks about using the local display for unraid, the GTX for the VM. At about 6:05 he talks about "planning to use the unraid desktop GUI, makes sure you plug your monitor into the onboard graphics". Then at 6:32 you see the boot and what is definitely some sort of desktop gui login of some sort. Again at 16:20 or so, you see more of this unraid desktop gui logon screen.
May 7, 20188 yr Doesn't the Storinator 45 use a RocketRAID 750 card which has the Marvell 88SE9485 chipset? Marvell chipsets cause a parity bug with unRAID... was this an issue with the Storinator?
May 8, 20188 yr 2 and a half years later... RocketRAID cards use very specialised firmware and software - they're not just a Marvell chipset with a different label.
Archived
This topic is now archived and is closed to further replies.