LinuxQuestions.org
Download your favorite Linux distribution at LQ ISO.
Home Forums Tutorials Articles Register
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - General
User Name
Password
Linux - General This Linux forum is for general Linux questions and discussion.
If it is Linux Related and doesn't seem to fit in any other forum then this is the place.

Notices


Reply
  Search this Thread
Old 10-01-2004, 04:53 PM   #1
BrianK
Senior Member
 
Registered: Mar 2002
Location: Los Angeles, CA
Distribution: Debian, Ubuntu
Posts: 1,334

Rep: Reputation: 51
100% of RAM in use all the time


My file server's RAM is alawys completely full (as far as I can tell from top). If I boot it up, the ram is, for the most part, not utilitized, but as soon as NFS clients start writing to it, it fills up. It works fine after that and never really goes into swap, but the NFS read/write performance goes down the tubes once the RAM is full.

Is there a reason why this is happening?

My file server is running RH9 with the 2.4.20-8smp kernel. I can't upgrade the kernel as it is the only kernel my crummy SATA RAID controller works under.

Thanks for any help.

FYI: from top:
Code:
Mem:  1029524k av, 1017112k used,   12412k free,       0k shrd,   78628k buff
                    853220k actv,   48332k in_d,   20508k in_c
Swap: 2040244k av,   43208k used, 1997036k free                  859388k cached
 
Old 10-01-2004, 04:56 PM   #2
ekaqu
Member
 
Registered: Sep 2004
Distribution: Debian SID, Debian Woody, Debian Sarge
Posts: 39

Rep: Reputation: 15
I have the same prob with my computer, "top" and "free meminfo" show that all my ram is filled. but when i use something like gkrellm, it shows how much ram i am really using
 
Old 10-01-2004, 05:07 PM   #3
amosf
Senior Member
 
Registered: Jun 2004
Location: Australia
Distribution: Mandriva/Slack - KDE
Posts: 1,672

Rep: Reputation: 46
Need to see the results of the free command to see how much ram is actually tied up...
 
Old 10-01-2004, 05:21 PM   #4
BrianK
Senior Member
 
Registered: Mar 2002
Location: Los Angeles, CA
Distribution: Debian, Ubuntu
Posts: 1,334

Original Poster
Rep: Reputation: 51
Quote:
Originally posted by amosf
Need to see the results of the free command to see how much ram is actually tied up...
ahh... didn't know about free, but the results are the same.
Code:
$ free
             total       used       free     shared    buffers     cached
Mem:       1029524     957912      71612          0      80540     797296
-/+ buffers/cache:      80076     949448
Swap:      2040244      44808    1995436
 
Old 10-01-2004, 05:46 PM   #5
amosf
Senior Member
 
Registered: Jun 2004
Location: Australia
Distribution: Mandriva/Slack - KDE
Posts: 1,672

Rep: Reputation: 46
According to free you only have 80 meg of the ram actually tied up in apps with 949meg being used as cache and buffers. This is normal for linux. It will use any 'free' memory for cache rather than waste it... the free value minus cache/buffers is a more reliable way to judge 'free' memory... ie in this case it's 949448 'free'...

The NFS problem might be due to something else?

And is there no way to use your SATA Raid with a (better) 2.6 kernel? What raid is it?
 
Old 10-01-2004, 06:08 PM   #6
BrianK
Senior Member
 
Registered: Mar 2002
Location: Los Angeles, CA
Distribution: Debian, Ubuntu
Posts: 1,334

Original Poster
Rep: Reputation: 51
Quote:
Originally posted by amosf
According to free you only have 80 meg of the ram actually tied up in apps with 949meg being used as cache and buffers. This is normal for linux. It will use any 'free' memory for cache rather than waste it... the free value minus cache/buffers is a more reliable way to judge 'free' memory... ie in this case it's 949448 'free'...
hmm.. ok. I suppose I could have suspected as much, however, it still seems odd that NFS suffers at the point at which all the RAM is being used for cache.
Quote:

The NFS problem might be due to something else?

And is there no way to use your SATA Raid with a (better) 2.6 kernel? What raid is it?
Promise FastTrack 150 SX4, running RAID 5.

The Promise website is down at the moment, but at the time I installed the card (February?) the 2.4 kernel was the only one supported - the 2.4.20-8 was listed explicitly in the instructions as being the *only* kernel this card would run under (and happens to be the kernel the distro came with). I had to call tech support a few times during the install & every time I called, they made sure I was running 2.4.20-8. I guess you get what you pay for.
 
Old 10-01-2004, 07:23 PM   #7
amosf
Senior Member
 
Registered: Jun 2004
Location: Australia
Distribution: Mandriva/Slack - KDE
Posts: 1,672

Rep: Reputation: 46
"hmm.. ok. I suppose I could have suspected as much, however, it still seems odd that NFS suffers at the point at which all the RAM is being used for cache."

Might have something to do with the way it's caching the data... Don't know

"Promise FastTrack 150 SX4, running RAID 5."

I think there is native support for this in the latest 2.6 kernels, but since it has software raid components then the binary proprietry promise driver you are using will be faster at this stage I'd say...

Not sure what the problem is, but the memory usage seems okay at least.
 
Old 10-01-2004, 09:15 PM   #8
trickykid
LQ Guru
 
Registered: Jan 2001
Posts: 24,149

Rep: Reputation: 269Reputation: 269Reputation: 269
Using RAM is good. If you never use it all, why have that much to begin with? This is something Windows puts to waste and is good at.
 
Old 10-04-2004, 03:00 PM   #9
BrianK
Senior Member
 
Registered: Mar 2002
Location: Los Angeles, CA
Distribution: Debian, Ubuntu
Posts: 1,334

Original Poster
Rep: Reputation: 51
Quote:
Originally posted by trickykid
Using RAM is good. If you never use it all, why have that much to begin with? This is something Windows puts to waste and is good at.
well, if nothing was effected by the RAM being in use, I'd agree. However, the fact that my NFS performance drops when it is completely in use makes me wish that it wasn't.
 
Old 10-05-2004, 07:27 AM   #10
trickykid
LQ Guru
 
Registered: Jan 2001
Posts: 24,149

Rep: Reputation: 269Reputation: 269Reputation: 269
Quote:
Originally posted by BrianK
well, if nothing was effected by the RAM being in use, I'd agree. However, the fact that my NFS performance drops when it is completely in use makes me wish that it wasn't.
Then I'd suggest running Memory tests then, if its not swapping but once its full its just degrading performance, that points to bad RAM or memory leaks, etc.
 
Old 10-05-2004, 02:05 PM   #11
BrianK
Senior Member
 
Registered: Mar 2002
Location: Los Angeles, CA
Distribution: Debian, Ubuntu
Posts: 1,334

Original Poster
Rep: Reputation: 51
Quote:
Originally posted by trickykid
Then I'd suggest running Memory tests then, if its not swapping but once its full its just degrading performance, that points to bad RAM or memory leaks, etc.
what has come up in another thread is that my RAID controller is software based w/ hardware acceleration. I'm guessing this has something to do with it & seems to make the most sense.

If it was an issue with bad memory, I'd expect that I'd be getting errors more than slowness. If it was a memory leak, I'd expect to run into swap space once the mem is full, which it isn't. Looks like I'll chalk this one up as a cheap controller card & get a new one once I get another terabyte of space to back up my current data.
 
Old 11-02-2004, 09:52 AM   #12
bono32
LQ Newbie
 
Registered: Nov 2004
Posts: 1

Rep: Reputation: 0
Question memory leak or something else?!

I have the same problem with the same SMP : 2.4.80-8SMP RedHat 9.
My memory gets full in couple of hours, but even if I shut down every application, the usage drops only 10-25%. I think it clearly indicates a memory leak but I was unable to detect what is the source. I suspect GNOME or X but if it was the case it would have been detected or reported by others earlier, no?
And something more, after few days of running I am unable to start a new Mozilla window or any Office window, it crashes and eventually I have to reboot my "solid" Linux Box.

Anyway, help appreciated...

I was going to paste in some data, but I had to reboot this machine few minutes ago. I will do it later.
 
  


Reply



Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
Crond Taking Almost 100% Processor Time ngwasuma Linux - Software 4 01-04-2006 09:05 AM
Distribution Suggestion for 100 MHz, 40 mb Ram whit2333 Linux - Distributions 5 06-27-2005 01:51 PM
100+ posts - time to say hello pongmaster LinuxQuestions.org Member Intro 1 08-27-2004 04:43 PM
Guess the Correct Time LQ will hit 100,000 Members jeremy LQ Suggestions & Feedback 152 05-03-2004 12:56 AM
Grip - rips are very slow and take 100% cpu time Mad Merlin Linux - Software 3 05-02-2004 12:16 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - General

All times are GMT -5. The time now is 12:14 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration