Jump to content

JorgeB

Moderators
  • Posts

    67,112
  • Joined

  • Last visited

  • Days Won

    703

Everything posted by JorgeB

  1. No need to clear the data as long as it fits in the new profile, though it will take more time depending on how much data there is and backing up anything important before starting is always a good idea.
  2. Me too, but currently at 99TB on my largest server (and same for the twin backup server).
  3. Seems like a disk issue, though you might want to try again after updating the LSI firmware to latest, you're on 15.00.00.00, current one is 20.00.07.00
  4. Only some SAS devices have dual port, if you're using SATA only one will work.
  5. IMHO it's the right thing to do, Marvell based controllers are nothing but trouble since v6, though some users still use them without major issues, I have four SASLP and two SAS2LP that I bough new a few years ago, over 100€ each, now they are only good for the bracket, when some of the LSIs I bough came with low profile bracket I can used those since they are the same.
  6. SFF-8643 to SFF-8643 from the HBA to the expander. SFF-8643 to SFF-8087 from the expander to the backplane.
  7. Make sure you're using the right cables, you need forward breakout cables, reverse breakout look the same but won't work for this.
  8. You can slow or even stop some fans, but yeah, some times it's difficult for home users to keep them in the optimal operating temps.
  9. Yes, low temps can be as bad as high temps, there a google study about it, you can google it
  10. That's a raid controller, though it might be flashable to IT mode, if it isn't it's not a good option, 9207 is IT mode only.
  11. Likely this, though it might take a while, since this issues aren't new.
  12. One more data point, it doesn't happen to me with btrfs disks, only with xfs.
  13. It's been like that for years, as long as I can remember, and it's not enormous, it's 20% free RAM, though that can be a considerable size for servers with large amounts of RAM.
  14. That's normal, by default 20% free RAM is used for write cache.
  15. Yes, but with a SAS2 PCIe 3.0 HBA it still has 4400MB/s usable total, so around 157MB/ per disks if fully loaded, a SAS3 HBA + expander would give you around 6000MB/s total. Yes, if you can afford it, just divided 4400MB/s (if using the same PCIe 3.0 SAS2 HBAs) by the number of disks connected to see how much more. Yes, but only if used with a SAS3 expander from LSI. Edit: See here for some real worlds tests:
  16. I can now reproduce this, the missing piece was that for me at least, it only starts happening after a large transfer, i.e., after a fresh boot on my test server I start seeing the reads on all disks at around the 30% mark of a 30GB transfer, and it's easily reproducible after a reboot, with v6.6.6 I can complete the same transfer without the reads, so maybe this can help LT do the same. P.S.: test server has 8GB of RAM, since the RAM amount can likely affect if/when this happens.
  17. It's a kernel issue, and it's not the first time with Marvell controllers.
  18. An 8 port HBA would be enough, preferably a PCIe 3.0 one, like 9207-8i, both miniSAS ports connect to the expander and the expander connects to the backplane, up to 28 devices total, together with the existing 9201-16i you can connect up to 44 devices.
  19. Those are loop devices, there's no problem with those.
  20. SAS3 HBA + SAS3 LSI expander would give you a little more bandwidth, but not that much more and it's much more expensive, you'll be fine with SAS2.
  21. This might help explain why I can't find many other Linux users complain about the LSI trim issue, if they are using ext4 looks like it still works, still would expect at least some xfs users complaining, maybe not so much with btrfs since it's likely much less used.
  22. Diags from v6.7, when there's a problem, would be the ones needed.
×
×
  • Create New...