4 ms·

That's not useful as a rule of thumb, since you can't know the size of "all inactive anonymous pages" without doing extensive runtime analysis of the system und

by tremon 9mo ago

That's not useful as a rule of thumb, since you can't know the size of "all inactive anonymous pages" without doing extensive runtime analysis of the system under consideration. That's pretty much the opposite of what a rule of thumb is for.

man8alexd 9mo ago
You are right, it is not a rule of thumb, and you can't determine optimal swap size right away. But you don't need "extensive runtime analysis". Start with a small swap - a few hundred megabytes (assuming the system has GBs of RAM). Check its utilization periodically. If it is full, add a few hundred megabytes more. That's all.
- ZoomZoomZoom 9mo ago
  It's not like it's easy to shuffle partitions around. Swap files are a pain, so you need to reserve space at the end of the table. By the time you need to increase swap the previous partition is going to be full. Better overcommit right away and live with the feeling you're wasting space.
  - man8alexd 9mo ago
    Exactly opposite. Don't use swap partitions, and use swap files, even multiple if necessary. Never allocate too much swap space. It is better to get OOM earlier then to wait for unresponsive system.
    - ZoomZoomZoom 9mo ago
      Swap partition is set and forget. Can be detected by label automatically, never fails. Swap file means fallocating, setting extended attributes (like `nocow`), finding file offset and writing it to kernel params, and other gotchas, like btrfs not allowing snapshotting a subvolume with an active swap file. Technically it's preferable, won't argue with that.
  - direwolf20 9mo ago
    Hast thou discovered our lord and savior LVM?
  - rascul 9mo ago
    > Swap files are a pain Easier than partitions: mkswap --size 2G --file swap.img swapon swap.img
    - ZoomZoomZoom 9mo ago
      Yeah, until you need to hibernate to one. I understand that calculating file offsets is not rocket science, but still, all the dance required is not exactly uninvolved and feels a bit fragile.

by tremon 9mo ago

man8alexd 9mo ago

You are right, it is not a rule of thumb, and you can't determine optimal swap size right away. But you don't need "extensive runtime analysis". Start with a small swap - a few hundred megabytes (assuming the system has GBs of RAM). Check its utilization periodically. If it is full, add a few hundred megabytes more. That's all.

ZoomZoomZoom 9mo ago
It's not like it's easy to shuffle partitions around. Swap files are a pain, so you need to reserve space at the end of the table. By the time you need to increase swap the previous partition is going to be full. Better overcommit right away and live with the feeling you're wasting space.
- man8alexd 9mo ago
  Exactly opposite. Don't use swap partitions, and use swap files, even multiple if necessary. Never allocate too much swap space. It is better to get OOM earlier then to wait for unresponsive system.
  - ZoomZoomZoom 9mo ago
    Swap partition is set and forget. Can be detected by label automatically, never fails. Swap file means fallocating, setting extended attributes (like `nocow`), finding file offset and writing it to kernel params, and other gotchas, like btrfs not allowing snapshotting a subvolume with an active swap file. Technically it's preferable, won't argue with that.
- direwolf20 9mo ago
  Hast thou discovered our lord and savior LVM?
- rascul 9mo ago
  > Swap files are a pain Easier than partitions: mkswap --size 2G --file swap.img swapon swap.img
  - ZoomZoomZoom 9mo ago
    Yeah, until you need to hibernate to one. I understand that calculating file offsets is not rocket science, but still, all the dance required is not exactly uninvolved and feels a bit fragile.