4 ms·

So... running Linux swapless is a thing? How popular is it?

by makz 7y ago

sandov 7y ago
I used to disable swap because it supposedly reduced the life span of SSDs, but I don't care about that anymore — when it dies, it dies.
yongjik 7y ago
Kubernetes, for example, doesn't even support swap. Some bug reports say it won't even run with swap enabled, though I didn't test myself. ¯\_(ツ)_/¯
- jillesvangurp 7y ago
  I don't know about that but it is very common to specify CPU and memory limits for docker containers. Exceeding those automatically leads to the process being killed. The reasoning is very simple: any form of swapping is completely unacceptable on a production server because it randomly and massively degrades server performance. If you have a cluster of stuff and one node is misbehaving like that, you kill it because that is completely unacceptable. If that is a regular thing, your servers are obviously mis-configured in some way and you fix it by provisioning more hardware or tweaking the limits. 12 years ago, before I got a mac, I had a windows XP laptop with enough memory (8GB) to disable the swap file (which world+dog will insist is a very stupid thing to do). This was great and vastly extended the useful life of my laptop. Alt+tabs were instant and I could run e.g. JVM applications with sane heap settings as well as a browser, office stuff and a few small things I needed with zero issues. On the rare occasion that something did run out of memory, it died or I killed it. Laptop disks were stupendously slow at the time; any form of swapping on a slow laptop disk is extremely disruptive. SSDs are much better but there too it tends to be mostly redundant. IMHO most forms of swapping are highly undesirable on both servers and end user hardware. Swapping to free up memory for file caching is simply unacceptable when you can instead just evict cache pages. If you don't have enough memory left to cache effectively, that just means things like memory mapped files will get a lot slower. If something allocates more memory than you have just kill it.
  - toast0 7y ago
    > The reasoning is very simple: any form of swapping is completely unacceptable on a production server because it randomly and massively degrades server performance. If you have a cluster of stuff and one node is misbehaving like that, you kill it because that is completely unacceptable. If that is a regular thing, your servers are obviously mis-configured in some way and you fix it by provisioning more hardware or tweaking the limits. A small swap space (~ 1G on a 64+G ram server) is a reasonable backstop against a slow memory leak. Assuming you don't have filesystem pages evicting anonymous pages, swap use is a clear indicator of too much memory use and points you in the direction of something to fix; and gives you a little bit of time to fix it on the running system. As long as swap is very small relative to ram, it's not going to enable thrashing -- a big leak or burst in use isn't going to fit in swap and you're going to be dead anyway.
  - Dylan16807 7y ago
    > If you don't have enough memory left to cache effectively, that just means things like memory mapped files will get a lot slower. Understatement of the year. If you can't cache your memory mapped executable pages, your system will be just as slow as in the nightmare swap scenario.
  - djsumdog 7y ago
    On old Win95 machines, you couldn't disable swap, because DirectX used it to buffer audio in several games (Jedi Knight comes to mind) and so your game wouldn't have any sound with virtual memory disabled.
- pritambarhate 7y ago
  Yes, here is the link to the issue: https://github.com/kubernetes/kubernetes/issues/53533 https://github.com/kubernetes/kubernetes/issues/53533 Still open.
kokey 7y ago
I don't know what other people do, but I think the better option would be to set vm.swappiness to 0. Swap space is a good safety valve. You should never really have to use it, so a good way to detect that something is going really wrong and take action before it brings the system is down is by looking at when swap is busy filling up. Also if someone opens up an application that grabs huge chunks of RAM but leave a lot of it idle, and turn swap off completely, they should not be surprised. I don't know why people see this as a bug, but perhaps I've just been spending time in the UNIX family tree for too many decades.
snazz 7y ago
No idea how popular it is, but that’s how I run my ThinkPad X220 with 4 GB RAM and a mechanical hard drive. It’s still incredibly snappy.

So... running Linux swapless is a thing? How popular is it?

by makz 7y ago

So... running Linux swapless is a thing? How popular is it?

I used to disable swap because it supposedly reduced the life span of SSDs, but I don't care about that anymore — when it dies, it dies.

yongjik 7y ago

Kubernetes, for example, doesn't even support swap. Some bug reports say it won't even run with swap enabled, though I didn't test myself. ¯\_(ツ)_/¯

jillesvangurp 7y ago
I don't know about that but it is very common to specify CPU and memory limits for docker containers. Exceeding those automatically leads to the process being killed. The reasoning is very simple: any form of swapping is completely unacceptable on a production server because it randomly and massively degrades server performance. If you have a cluster of stuff and one node is misbehaving like that, you kill it because that is completely unacceptable. If that is a regular thing, your servers are obviously mis-configured in some way and you fix it by provisioning more hardware or tweaking the limits. 12 years ago, before I got a mac, I had a windows XP laptop with enough memory (8GB) to disable the swap file (which world+dog will insist is a very stupid thing to do). This was great and vastly extended the useful life of my laptop. Alt+tabs were instant and I could run e.g. JVM applications with sane heap settings as well as a browser, office stuff and a few small things I needed with zero issues. On the rare occasion that something did run out of memory, it died or I killed it. Laptop disks were stupendously slow at the time; any form of swapping on a slow laptop disk is extremely disruptive. SSDs are much better but there too it tends to be mostly redundant. IMHO most forms of swapping are highly undesirable on both servers and end user hardware. Swapping to free up memory for file caching is simply unacceptable when you can instead just evict cache pages. If you don't have enough memory left to cache effectively, that just means things like memory mapped files will get a lot slower. If something allocates more memory than you have just kill it.
- toast0 7y ago
  > The reasoning is very simple: any form of swapping is completely unacceptable on a production server because it randomly and massively degrades server performance. If you have a cluster of stuff and one node is misbehaving like that, you kill it because that is completely unacceptable. If that is a regular thing, your servers are obviously mis-configured in some way and you fix it by provisioning more hardware or tweaking the limits. A small swap space (~ 1G on a 64+G ram server) is a reasonable backstop against a slow memory leak. Assuming you don't have filesystem pages evicting anonymous pages, swap use is a clear indicator of too much memory use and points you in the direction of something to fix; and gives you a little bit of time to fix it on the running system. As long as swap is very small relative to ram, it's not going to enable thrashing -- a big leak or burst in use isn't going to fit in swap and you're going to be dead anyway.
- Dylan16807 7y ago
  > If you don't have enough memory left to cache effectively, that just means things like memory mapped files will get a lot slower. Understatement of the year. If you can't cache your memory mapped executable pages, your system will be just as slow as in the nightmare swap scenario.
- djsumdog 7y ago
  On old Win95 machines, you couldn't disable swap, because DirectX used it to buffer audio in several games (Jedi Knight comes to mind) and so your game wouldn't have any sound with virtual memory disabled.
pritambarhate 7y ago
Yes, here is the link to the issue: https://github.com/kubernetes/kubernetes/issues/53533 https://github.com/kubernetes/kubernetes/issues/53533 Still open.

kokey 7y ago

I don't know what other people do, but I think the better option would be to set vm.swappiness to 0. Swap space is a good safety valve. You should never really have to use it, so a good way to detect that something is going really wrong and take action before it brings the system is down is by looking at when swap is busy filling up. Also if someone opens up an application that grabs huge chunks of RAM but leave a lot of it idle, and turn swap off completely, they should not be surprised. I don't know why people see this as a bug, but perhaps I've just been spending time in the UNIX family tree for too many decades.

snazz 7y ago

No idea how popular it is, but that’s how I run my ThinkPad X220 with 4 GB RAM and a mechanical hard drive. It’s still incredibly snappy.