Re: tmpfs panic

From: Kostik Belousov <kostikbel_at_gmail.com>
Date: Sun, 15 Jun 2008 14:12:52 +0300
On Sun, Jun 15, 2008 at 03:05:45PM +0400, Dmitry Morozovsky wrote:
> On Sun, 15 Jun 2008, Kostik Belousov wrote:
> 
> KB> On Sun, Jun 15, 2008 at 01:06:24PM +0400, Dmitry Morozovsky wrote:
> KB> > Hi there,
> KB> > 
> KB> > at contemporary RELENG_7/amd64
> 
> [snip]
> 
> 
> KB> I suspect this may be my mistake.
> KB> IN case you can reproduce it, please, try the patch below.
> 
> Will try, thank you. 
> 
> KB> > Also, active tmpfs usage easy leads to "swap zone exhausted, increase
> KB> > kern.maxswzone", even with 2G RAM + 4G swap and approx 2-3G of tmpfs
> KB> > in use -- any hints?
> KB> 
> KB> I think the message is pretty much self-explanatory. Kernel tried to
> KB> allocate the meatadata to track the swap metadata, and zone appears
> KB> exhausted. It is not the swap space shortage. Instead, this is kernel
> KB> zone used to track swap allocation shortage.
> KB> 
> KB> It is quite non-obvious how to automatically tune this limit, since zone
> KB> is allocated before swap is configured.
> 
> Well, but loader(8) man page said:
>   The default of 32MB allows the kernel to support a maximum of ~7GB of swap. 
> 
> and I have only 4G, and system freezes with maxswzone exaustion well before 
> that...
I suspect that the estimation both outdated and arch-specific (i.e. at
least divide it by two for amd64).

Received on Sun Jun 15 2008 - 09:12:57 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:32 UTC