home.social

#openzfsmastery — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #openzfsmastery, aggregated by home.social.

fetched live
  1. Today in #openzfsmastery: going through @allanjude's comments and putting them in place in the manuscript.

    It's not enough to just make words. Gotta make them in the right place.

    Want to help me keep the rats fed as I work on this? mwl.io/sponsor

  2. Today in #openzfsmastery: going through @allanjude's comments and putting them in place in the manuscript.

    It's not enough to just make words. Gotta make them in the right place.

    Want to help me keep the rats fed as I work on this? mwl.io/sponsor

  3. Today in #openzfsmastery

    "Just because the pool has larger disks, doesn’t necessarily mean that you want to expand the pool. Perhaps those larger hard drives are destined for a separate pool, but you’re swapping them in to cover a mechanical failure. Maybe you needed to get these installed for scheduling reasons, but the database team hasn’t yet signed the paper where they officially agree to pay for the drives."

    I'd appreciate your help writing this book. mwl.io/sponsor

  4. Today in #openzfsmastery

    "Just because the pool has larger disks, doesn’t necessarily mean that you want to expand the pool. Perhaps those larger hard drives are destined for a separate pool, but you’re swapping them in to cover a mechanical failure. Maybe you needed to get these installed for scheduling reasons, but the database team hasn’t yet signed the paper where they officially agree to pay for the drives."

    I'd appreciate your help writing this book. mwl.io/sponsor

  5. "Yes, you could replace all the disks simultaneously, if you’d like to shatter your day with a glorious festival of I/O contention." #openzfsmastery

    Wrote about log and cache devices today, and expanding pools by replacing disks with larger disks.

    The book is underway! I'd appreciate your help keeping the lights on as I finish it. mwl.io/sponsor

  6. "Yes, you could replace all the disks simultaneously, if you’d like to shatter your day with a glorious festival of I/O contention." #openzfsmastery

    Wrote about log and cache devices today, and expanding pools by replacing disks with larger disks.

    The book is underway! I'd appreciate your help keeping the lights on as I finish it. mwl.io/sponsor

  7. Today in #openzfsmastery

    'The computational engineering team assured you that the data on the pool named “disposable” was disposable. They swore they would never put anything important on it. Three weeks later, someone called to request you restore a vital file. When you said you didn’t back up the disposable share and that the data there was completely unprotected, they informed you that the “disposable” files were, in fact, essential.

    It would be maddening, if before that first meeting you hadn’t predicted exactly this. To your supervisor. In writing.'

  8. Today in #openzfsmastery

    'The computational engineering team assured you that the data on the pool named “disposable” was disposable. They swore they would never put anything important on it. Three weeks later, someone called to request you restore a vital file. When you said you didn’t back up the disposable share and that the data there was completely unprotected, they informed you that the “disposable” files were, in fact, essential.

    It would be maddening, if before that first meeting you hadn’t predicted exactly this. To your supervisor. In writing.'

  9. dived into Linux kernel modules today and got some missing tidbits of #openzfsmastery filled in so @allanjude can take a look.

    But I also got to use the phrase "Oracle taint," so the day wasn't a total loss.

    I'd appreciate your support writing this book. mwl.io/sponsor

  10. dived into Linux kernel modules today and got some missing tidbits of #openzfsmastery filled in so @allanjude can take a look.

    But I also got to use the phrase "Oracle taint," so the day wasn't a total loss.

    I'd appreciate your support writing this book. mwl.io/sponsor

  11. "Oracle casts a noxious cloud over technology and their presence in open source is a toxic taint."

    Torn between "too far" and "make sure this sentence retains ambiguity over which definition of 'taint' I'm using." #openzfsmastery

  12. "Oracle casts a noxious cloud over technology and their presence in open source is a toxic taint."

    Torn between "too far" and "make sure this sentence retains ambiguity over which definition of 'taint' I'm using." #openzfsmastery

  13. Today in #openzfsmastery: replacing failed drive s0d1-NCC-1701D with new drive s0d5-NCC-1701E.

    Yes, I'm easily amused. Or desperate for amusement. One of those.

    Anyway, you could sponsor this thing and help me feed the rats while I finish writing it. mwl.io/sponsor

  14. Today in #openzfsmastery: replacing failed drive s0d1-NCC-1701D with new drive s0d5-NCC-1701E.

    Yes, I'm easily amused. Or desperate for amusement. One of those.

    Anyway, you could sponsor this thing and help me feed the rats while I finish writing it. mwl.io/sponsor

  15. Huh. Causing #openZFS failures is much more difficult now than it was ten years ago. #openzfsmastery

  16. Huh. Causing #openZFS failures is much more difficult now than it was ten years ago. #openzfsmastery

  17. "Lucas has run high-volume netflow collectors, where data is written all day every day for as long as the doomed disks endure. He considered adjusting metadata redundancy when the disk couldn’t keep up. If it helped, he copied new data off to more fortunate storage every hour and started budgeting for faster storage."

    #openzfsmastery datasets chapter done at last! On to scrubs, repairs, and pool changes.

    Want your name in the book? Sponsors get that! mwl.io/sponsor

  18. "Lucas has run high-volume netflow collectors, where data is written all day every day for as long as the doomed disks endure. He considered adjusting metadata redundancy when the disk couldn’t keep up. If it helped, he copied new data off to more fortunate storage every hour and started budgeting for faster storage."

    #openzfsmastery datasets chapter done at last! On to scrubs, repairs, and pool changes.

    Want your name in the book? Sponsors get that! mwl.io/sponsor

  19. Today in #openzfsmastery

    "This is one of those times when the manual's dire warning is insufficient."

  20. Today in #openzfsmastery

    "This is one of those times when the manual's dire warning is insufficient."

  21. Today's #openzfsmastery typo: "zroot/ROOT/default is our rot filesystem."

    Inarguable.

    I'd appreciate your help surviving as I finish writing this book. mwl.io/sponsor

  22. Today's #openzfsmastery typo: "zroot/ROOT/default is our rot filesystem."

    Inarguable.

    I'd appreciate your help surviving as I finish writing this book. mwl.io/sponsor

  23. "Properly written setuid programs are not risky. That's why real setuid programs are risky."

    Progress on #openzfsmastery today! Hoping to finish the dataset chapter tomorrow. (Great. Now I've jinxed it.)

    Want to help me survive writing this? mwl.io/sponsor

  24. "Properly written setuid programs are not risky. That's why real setuid programs are risky."

    Progress on #openzfsmastery today! Hoping to finish the dataset chapter tomorrow. (Great. Now I've jinxed it.)

    Want to help me survive writing this? mwl.io/sponsor

  25. Churning through the #openzfsmastery
    dataset chapter. Progress happened today!

    I'd appreciate your support in making this book happen. mwl.io/sponsor

  26. Churning through the #openzfsmastery
    dataset chapter. Progress happened today!

    I'd appreciate your support in making this book happen. mwl.io/sponsor

  27. That's the #openzfsmastery "pools" chapter done!

    Want to help me survive writing this book? Want your name in the book so you get your share of blame? Sponsor it at mwl.io/sponsor

  28. That's the #openzfsmastery "pools" chapter done!

    Want to help me survive writing this book? Want your name in the book so you get your share of blame? Sponsor it at mwl.io/sponsor

  29. RE: io.mwl.io/@mwl/116822660826509

    TFW it's been a while and @mwl mentions working on the upcoming #openzfsmastery and that sponsorships are open, but Captain Memory-Span-of-a-Goldfish has forgotten whether he already sponsored it.

    $ grep MWL transactions.ledger | grep -c openzfs
    1

    Phew. Sponsored.

  30. RE: io.mwl.io/@mwl/116822660826509

    TFW it's been a while and @mwl mentions working on the upcoming #openzfsmastery and that sponsorships are open, but Captain Memory-Span-of-a-Goldfish has forgotten whether he already sponsored it.

    $ grep MWL transactions.ledger | grep -c openzfs
    1

    Phew. Sponsored.

  31. working on #openzfsmastery performance vs resilience section.

    Thinking that the fault tolerance of a 3-disk striped VDEV can best be described as "yeet." #sysadmin

    This beast is open for sponsorship. mwl.io/sponsor

  32. working on #openzfsmastery performance vs resilience section.

    Thinking that the fault tolerance of a 3-disk striped VDEV can best be described as "yeet." #sysadmin

    This beast is open for sponsorship. mwl.io/sponsor

  33. Got into the #openzfsmastery manuscript and found I had to fact-check a Debian detail.

    Which means getting my crashbox online.

    My office has no cabling. Connection options are a wifi bridge or ethernet-over-power.

    If I use ether-over-power, I must either daisy-chain powerstrips or buy a powerstrip with many plugs and a ten-foot cord. My experience with EoP has been unpleasant, but that might have been because the old house was wired in 1949 by a clearly PTSD-addled pipefitter.

    I have on hand a couple of Mikrotiks with wifi. In theory, a bridge should be easy. In practice, I'll probably burn the rest of the day on this. :flan_despair:

  34. Got into the #openzfsmastery manuscript and found I had to fact-check a Debian detail.

    Which means getting my crashbox online.

    My office has no cabling. Connection options are a wifi bridge or ethernet-over-power.

    If I use ether-over-power, I must either daisy-chain powerstrips or buy a powerstrip with many plugs and a ten-foot cord. My experience with EoP has been unpleasant, but that might have been because the old house was wired in 1949 by a clearly PTSD-addled pipefitter.

    I have on hand a couple of Mikrotiks with wifi. In theory, a bridge should be easy. In practice, I'll probably burn the rest of the day on this. :flan_despair:

  35. Going offline to drag some more #openzfsmastery back into my brain.

    I can't say that I'm back at work properly, but I'm headed in that direction. It feels good.

  36. Going offline to drag some more #openzfsmastery back into my brain.

    I can't say that I'm back at work properly, but I'm headed in that direction. It feels good.

  37. Chapters 0 and 1 of #openzfsmastery reviewed. Tweaked a few things but not too bad. Recorded a snippet for tomorrow's #60SecondsOfWIP, including a footnote.

    Want to help me with this book? Sponsor it at mwl.io/sponsor

  38. Chapters 0 and 1 of #openzfsmastery reviewed. Tweaked a few things but not too bad. Recorded a snippet for tomorrow's #60SecondsOfWIP, including a footnote.

    Want to help me with this book? Sponsor it at mwl.io/sponsor

  39. going offline to poke at the #openzfsmastery manuscript. It might eat me, wish me luck.

  40. going offline to poke at the #openzfsmastery manuscript. It might eat me, wish me luck.

  41. dear non-Patronizers: February's "See the Sausage Being Made" is now public. mwl.io/archives/24612

    includes a reveal of the #openzfsmastery cover painting!

  42. dear non-Patronizers: February's "See the Sausage Being Made" is now public. mwl.io/archives/24612

    includes a reveal of the #openzfsmastery cover painting!

  43. "They cranked the OpenZFS version up to 5000, leaving Oracle plenty of space to add new version numbers--as if Oracle would EVER innovate rather than merely exploit acquired intellectual property." #nonotryagain #openzfsmastery

  44. "They cranked the OpenZFS version up to 5000, leaving Oracle plenty of space to add new version numbers--as if Oracle would EVER innovate rather than merely exploit acquired intellectual property." #nonotryagain #openzfsmastery

  45. "Remember, happy filesystems are all alike; every unhappy filesystem is unhappy in its own way." #openzfsmastery

  46. "Remember, happy filesystems are all alike; every unhappy filesystem is unhappy in its own way." #openzfsmastery

  47. Ground out some words on #openzfsmastery today, discussing lopsided pools and self-healing. Also got tomorrow's #60SecondsOfWIP recorded!

    I'd appreciate your help keeping the lights on while I finish this book. sponsor.mwl.io

  48. Ground out some words on #openzfsmastery today, discussing lopsided pools and self-healing. Also got tomorrow's #60SecondsOfWIP recorded!

    I'd appreciate your help keeping the lights on while I finish this book. sponsor.mwl.io

  49. I should replace this whole section with "make sure your filesystem blocks align with your disk's physical sectors or ZFS will GUT YOU LIKE THE SQUEALING PIGLET YOU ARE." #openzfsmastery

  50. I should replace this whole section with "make sure your filesystem blocks align with your disk's physical sectors or ZFS will GUT YOU LIKE THE SQUEALING PIGLET YOU ARE." #openzfsmastery

  51. Today in #openzfsmastery: re-re-redrafted the section on block size, then discussed ashift autodetection, stripes, and mirrors.

    Want to help me keep the lights on while I scribble this silly thing? sponsor.mwl.io

  52. Today in #openzfsmastery: re-re-redrafted the section on block size, then discussed ashift autodetection, stripes, and mirrors.

    Want to help me keep the lights on while I scribble this silly thing? sponsor.mwl.io

  53. Finally getting back to my 72-disk draid bhyve VM for #openzfsmastery.

    Turns out that setting ahci_device_limit="8" in the vm-bhyve config let it boot just fine.

    Thank you everyone for your suggestions, but especially @dch for this one!

  54. Finally getting back to my 72-disk draid bhyve VM for #openzfsmastery.

    Turns out that setting ahci_device_limit="8" in the vm-bhyve config let it boot just fine.

    Thank you everyone for your suggestions, but especially @dch for this one!