50mb/s but now it's even worse. i rebooted and now i get this 21:34:12^root@datastorm:~/chunk > mdadm -A /dev/md0 /dev/sd[bcdefg] --backup-file=/root/chunk/1 mdadm: Failed to restore critical section for reshape, sorry. any ideas to force it? Alex On Thu, Dec 20, 2012 at 9:01 PM, Roger Heflin <rogerheflin@xxxxxxxxx> wrote: > On Thu, Dec 20, 2012 at 3:06 PM, Alex Pientka <alex.pientka@xxxxxxxxx> wrote: >> What kind of promise card and how is it connected to the motherboard? >> => pci-x promise 16300. Currently I have a ticket open with promise >> and I have been talking to the developer, who wrote the stex driver. >> He says it is a very lightweight driver. I assume it is a firmware >> issue on the promise card. If you check the promise release notes they >> constantly talk about hitachi 2tb drive issues all related to >> performance. My co-worker has 500gb, 750gb and 1000mb drives >> connected, which work fast. My 2tb drives are very slow. My writes are >> usually around 50mb/s. >> >> And how were the old disk controller connected to the motherboard? >> => abit pro ip35 motherboard; 5 of them connected to the intel on >> board controller and one disk was running on a sil sata2 pci-express >> 1x controller. This was necessary since I was not able to put the boot >> drive on the pci express controller. >> >> >> MB/s------------INTEL-SIL----PROMISE-- >> Writer Report--------584.91-----51.54--- >> Re-writer Report-----645.69----126.38--- >> Random Read Report--4190.72------6.12--- >> Random Write Report--474.83-----58.54--- >> Alex >> >> > > And the pci-x slot that it is in should be at least running at 66mhzx64bit so > should be able to do 500mb/second or so. > > I would do separate "dd if=/dev/<diskname> of=/dev/null bs=1M count=1000" on > the disks and see if that even gets good speed. > > If that gets good speed do all disks at once and see how fast they all > are at the same time, > "vmstat 1" (bi) will tell you how fast the whole group is running at > the same time. -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
