linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: keith <kmannth@us.ibm.com>
To: Hugh Dickins <hugh@veritas.com>
Cc: Andrew Morton <akpm@osdl.org>, linux-mm@kvack.org
Subject: Re: [Bug 3268] New: Lowmemory exhaustion problem with v2.6.8.1-mm4 16gb
Date: Wed, 25 Aug 2004 14:49:25 -0700	[thread overview]
Message-ID: <1093470564.5677.1920.camel@knk> (raw)
In-Reply-To: <Pine.LNX.4.44.0408252104540.2664-100000@localhost.localdomain>

On Wed, 2004-08-25 at 13:37, Hugh Dickins wrote:
> On Wed, 25 Aug 2004, keith wrote:
> > 
> > I turned CONFIG_DEBUG_PAGEALLOC off.
> 
> Good, thanks.
> 
> > I ran into problems when I tried
> > building 60 kernel trees or so.  It is using about 10gb of tmpfs space. 
> > See http://bugme.osdl.org/attachment.cgi?id=3566&action=view for more
> > info.
> 
> Okay, that's more to the point.
> 
> > It looks like there should be a maximum number of inodes for tmpfs.
> > Because I didn't know how many inodes it is using but I gave it a large
> > playground to use (mount -t tmpfs -o size=15G,nr_inodes=10000k,mode=0700
> > tmpfs /mytmpfs)
> 
> Well, by default it does give you a maximum number of inodes for tmpfs:
> which at 16GB of RAM (in 4KB pages: odd that that factors in, but I've
> no great urge to depart from existing defaults except where it's buggy)
> would give you 2M inodes.  But since each might roughly be expected to
> consume 1KB (a shmem_inode, a dentry, a longer name, a radix node) of
> low memory, we'd want 2GB of lowmem to support all those: too much.

I think there is a need for a hard upper limit to the number of inodes
tmpfs can use.  With this system I can have 32gb of memory and I also
have a 64gb(i386) system. tempfs sizes above 15=g are definitely
possible.  What about a 40gig tmpfs.... 
> 
> So, how well does the patch below work for you, if you leave nr_inodes
> to its default?  Carry on setting size=15G, that should be okay; but
> I don't think the kernel should stop you if you insist on setting
> nr_inodes to something unfortunate like 10M.  "df -i" will show
> how many inodes it's actually giving you by default.

I was wondering how to do that thanks. 
With the patch I only get 82294 inodes.  Hmmmm.... I don't have any
lowmem issues but I can't create too many files.  

Thanks,
  Keith Mannthey 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

  parent reply	other threads:[~2004-08-25 21:49 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1093400029.5677.1866.camel@knk>
2004-08-25 14:05 ` Hugh Dickins
2004-08-25 19:05   ` keith
2004-08-25 20:37     ` Hugh Dickins
2004-08-25 20:53       ` Andrew Morton
2004-08-25 20:59         ` Hugh Dickins
2004-08-25 21:19         ` Martin J. Bligh
2004-08-25 21:35           ` Andrew Morton
2004-08-25 21:49       ` keith [this message]
2004-08-25 22:02         ` Martin J. Bligh
2004-08-26 15:28         ` Hugh Dickins
     [not found] <200408242051.i7OKplP0009870@fire-1.osdl.org>
2004-08-24 21:43 ` Andrew Morton
2004-08-24 23:11   ` keith
2004-08-24 23:24     ` Andrew Morton
2004-08-24 23:20   ` Hugh Dickins

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1093470564.5677.1920.camel@knk \
    --to=kmannth@us.ibm.com \
    --cc=akpm@osdl.org \
    --cc=hugh@veritas.com \
    --cc=linux-mm@kvack.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox