Linux-Misc Digest #9, Volume #27 Sat, 3 Feb 01 11:13:03 EST
Contents:
Re: RAM advice ("Xavier Houppertz")
Re: .src.rpm Error "File Is Not A Regular File" In Redhat 7.0 (Jacob Mathew)
Re: emulate linux under a windows server? (Michael Heiming)
Re: Convert Word-DOC to PostScript ([EMAIL PROTECTED])
Re: CIFS and NFS file locking ("Peter T. Breuer")
Re: User write privileges to /dos partitions ("Peter T. Breuer")
Re: Modem [linmodem] configuration ("Peter T. Breuer")
Re: unapcking the 2.4 kernel in a home dir ("Peter T. Breuer")
Re: RAM advice (Jean-David Beyer)
Re: Modem [linmodem] configuration ("Panagiotis")
Re: User write privilege on /dos partition (Lee Webb)
Re: Calendar Software (Eric Ho)
Re: unapcking the 2.4 kernel in a home dir (Jean-David Beyer)
Re: RAM advice ("Xavier Houppertz")
----------------------------------------------------------------------------
From: "Xavier Houppertz" <[EMAIL PROTECTED]>
Subject: Re: RAM advice
Date: Sat, 3 Feb 2001 15:26:34 +0100
I still have 114 Mb but this is swapfile and thus much slower (but maybe
linux handle this so good ...)
I see that you have still 77 MB of free RAM, that's 'easy' !
My memory usage doesn't change too much so I'am always 'on the edge' and
when a user make a big query on the db, my server performance falls !
About kswapd, it is running (don't know what it is anyway). I only attached
the first page of the processes ...
Xavier.
"Jean-David Beyer" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Xavier Houppertz wrote:
> >
> > Hi,
> >
> > This is just for an advice. Do you think I'm 'easy' with this config or
> > should i add more memory ?
> >
> > 12:29pm up 72 days, 22:19, 11 users, load average: 0.25, 0.22, 0.16
> > 109 processes: 96 sleeping, 2 running, 11 zombie, 0 stopped
> > CPU states: 11.5% user, 1.3% system, 0.0% nice, 87.4% idle
> > Mem: 257088K av, 252764K used, 4324K free, 95864K shrd, 14028K buff
> > Swap: 130748K av, 8184K used, 122564K free 100220K
cached
> >
> You seem to have lots of memory. Recall that you have about 114
> megabytes of memory devoted to cache and buffers and this space can be
> used for programs when needed.
>
> > PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
> > 31052 dir 10 0 17580 17M 2196 S 0 7.4 6.8 0:41 java
> > 30890 dir 6 0 17864 17M 2204 R 0 3.5 6.9 0:53 java
> > 31307 root 6 0 748 748 564 R 0 1.5 0.2 0:02 top
> > 31377 root 5 0 276 276 232 S 0 0.1 0.1 0:00 sleep
> > 24946 dir 0 0 1296 1296 844 S 0 0.1 0.5 0:08 fvwm2
> > 1 root 0 0 336 296 276 S 0 0.0 0.1 0:01 init
> > 2 root 0 0 0 0 0 SW 0 0.0 0.0 0:05 kflushd
> > 3 root -12 -12 0 0 0 SW< 0 0.0 0.0 0:00 kswapd
> > 30880 root 0 0 10856 10M 1772 S 0 0.0 4.2 0:19 X
> > 444 root 0 0 832 768 640 S 0 0.0 0.2 0:00 bash
> > 270 sybase 0 0 5864 4628 780 S 0 0.0 1.8 1:10
> > backupserver
> > 29 root 0 0 292 256 240 S 0 0.0 0.0 0:00 kerneld
> > 211 bin 0 0 320 300 244 S 0 0.0 0.1 0:00 portmap
> > 225 root 0 0 400 372 320 S 0 0.0 0.1 0:09 syslogd
> > 234 root 0 0 516 344 304 S 0 0.0 0.1 0:00 klogd
> > 245 daemon 0 0 356 320 280 S 0 0.0 0.1 0:00 atd
> > 256 root 0 0 440 400 360 S 0 0.0 0.1 0:00 crond
> > 31511 root 5 0 736 736 540 S 0 0.0 0.2 0:13
> > printstock2
> > 264 sybase 2 0 29076 21M 22072 S 0 0.0 8.6 608:15
dataserver
> > 263 sybase 0 0 624 520 516 S 0 0.0 0.2 0:00
RUN_SYBASE
> > 290 root 0 0 1092 688 504 S 0 0.0 0.2 0:00 named
> > 268 sybase 0 0 612 508 504 S 0 0.0 0.1 0:00
> > RUN_SYB_BACK
> > 23638 root 0 0 1500 1492 1152 S 0 0.0 0.5 0:00 xdm
> > 315 root 0 0 540 520 428 S 0 0.0 0.2 0:00
rpc.mountd
> > 324 root 0 0 556 528 428 S 0 0.0 0.2 0:08
rpc.nfsd
> > 349 root 0 0 1060 1060 968 S 0 0.0 0.4 0:01 httpd
> > 338 root 0 0 284 268 236 S 0 0.0 0.1 0:01 gpm
>
> I am a little surprised by this list. Here is the top of mine:
>
> 7:47am up 5 days, 16:17, 3 users, load average: 2.05, 2.12, 2.09
> 72 processes: 69 sleeping, 3 running, 0 zombie, 0 stopped
> CPU0 states: 91.1% user, 8.1% system, 89.0% nice, 0.0% idle
> CPU1 states: 92.1% user, 7.1% system, 90.0% nice, 0.0% idle
> Mem: 516672K av, 444408K used, 72264K free, 57640K shrd, 41932K
> buff
> Swap: 273088K av, 16084K used, 257004K free 278796K
> cached
>
> PID USER PRI NI PAGEIN SIZE RSS SHARE STAT %CPU %MEM CTIME
> COMMAND
> 31621 seti 20 19 943 14876 14M 796 R N 45.3 2.8 619:13
> setiathome
> 31963 seti2 19 19 910 14864 14M 796 R N 44.4 2.8 585:55
> setiathome
> 32355 root 1 0 815 41044 40M 2080 S 5.0 7.9 4:40 X
> 945 jdbeyer 5 5 836 3504 3504 2736 S N 1.9 0.6 4:47
> cpumemusage_app
> 920 jdbeyer 10 8 369 916 916 700 R N 1.6 0.1 1:04 top
> 918 jdbeyer 0 0 393 1220 1220 1012 S 0.9 0.2 0:32
> xosview
> 885 jdbeyer 0 0 13805 3408 3408 1948 S 0.4 0.6 0:24
> enlightenment
> 1 root 0 0 403 284 268 208 S 0.0 0.0 14254m
> init
> 2 root 0 0 0 0 0 0 SW 0.0 0.0 0:04
> kflushd
> 3 root 0 0 0 0 0 0 SW 0.0 0.0 0:16
> kupdate
> 4 root 0 0 0 0 0 0 SW 0.0 0.0 0:00
> kpiod
> 5 root 0 0 4657 0 0 0 SW 0.0 0.0 0:05
> kswapd
> 6 root -20 -20 0 0 0 0 SW< 0.0 0.0 0:00
> mdrecoveryd
>
> I guess your distribution is very different from mine. I do not remember
> kerneld since Red Hat Linux 5.0 days. Where are your kpiod and kswapd
> processes? I suppose your distribution does not require them, but it is
> a surprise to see them missing.
>
> --
> .~. Jean-David Beyer Registered Linux User 85642.
> /V\ Registered Machine 73926.
> /( )\ Shrewsbury, New Jersey
> ^^-^^ 7:45am up 5 days, 16:14, 3 users, load average: 2.19, 2.16, 2.10
------------------------------
From: Jacob Mathew <[EMAIL PROTECTED]>
Crossposted-To: alt.linux,alt.os.linux
Subject: Re: .src.rpm Error "File Is Not A Regular File" In Redhat 7.0
Date: Sat, 3 Feb 2001 22:38:01 +0800
Reply-To: [EMAIL PROTECTED]
Hi there,
I'm completely new at this too, but if you're asking about how to install a
src.rpm, the following (unqualified) response, may help:
First, make sure that you have root privileges.
Second, make sure that you've followed the installation instructions
correctly, see below:
Basically, src.rpm files (aka source rpms or srpms) are packages, like
regular rpms, but ones that contain the source code of the program you are
trying to install, rather than the compiled programs themselves. Try the
following (I use Mandrake 7.2, so I'm not sure that the directories etc.
will be identical, but hopefully...):
1. Install the src.rpm package with:
rpm -i [name].src.rpm
2. change directory (cd) to /usr/src/RPM/SPECS/
3. Enter the following, to compile a binary package:
rpm -bb [name].spec
4. change directory to /usr/src/RPM/RPMS
5. Change to the directory for your architecture (e.g. i386, i486, k6 etc)
6. Here, you should (hopefully) find the built rpm. Install it as you would
a regular rpm, with:
rpm -i [name].[arch].rpm
(e.g. rpm -i ncftp-3.beta19-2.i686.rpm)
Thirdly, try downloading the package you wish to install again, as it may
be corrupt. (Try a different download site).
Failing all, see if you can find a pre-built rpm of the package you want to
install.
Hope this works (and isn't misleading).
Beret wrote:
> I have been trying to install many distros of Linux and I always seem
> to hit this same brick wall. I get the basics up and running,
> Internet access, sound, correct resolution etc but then I get
> frustrated, delete the partitions run fdisk/mbr and then go back to
> Windows for a while. Then I get restless, (I have been using and
> repairing Windows for almost 10 years) and try another time.
> -What is this brick wall, you are asking? I CAN"T RELIABLY INSTALL
> ANY APPS (excuse the shouting). I have been running around for months
> reading docs, how-to's, helps and mans to no avail. The ng's are my
> last resort.
> Now let's cut to the chase and get to the latest error message.
>
> I wanted to install the Redhat Errata RPM's for 7.0. and after a lot
> of struggling and researching I found
> http://www.buberel.org/linux/source-rpm.php I then realized that my
> command " rpm -i whatever.src.rpm" didn't do anything because those
> files, even though they have an extension of .rpm, they also have a
> "pre-extension???" of .src This means that they must be processed
> differently from .rpm's
>
> I followed the steps and when I got to
>
> rpm -Uvh ncftp-3.0beta19-2.i686.rpm
>
> I just got the following error: " File Is Not A Regular File"
>
> Also when I try and run the -i rpm command on regular .rpm's I also
> get an error which is roughly "can't install this file"
>
> I have been reading the ng's for a long time and some of you really
> know this black art that is Linux really well. If it weren't for you
> Linux would have gone the way of OS2 a long time ago. Keep up the
> good work :-)
>
>
------------------------------
Date: Sat, 03 Feb 2001 15:28:34 +0100
From: Michael Heiming <[EMAIL PROTECTED]>
Subject: Re: emulate linux under a windows server?
Harold Stevens ** PLEASE SEE SIG ** wrote:
> In <[EMAIL PROTECTED]>, Michael Heiming:
>
> [Snip...]
>
> |> Sure if someone does this he uses it the other way running M$ within Linux
> |> (VM-Ware), I've been told it's much more stable
> |> this way...
>
> [Snip...]
>
> Apparently, the American NSA agrees:
>
> http://www.zdii.com/industry_list.asp?mode=news&doc_id=ZD2681205
>
> I love the last line:
>
> In a nod to the open-source community, he said that--for the NSA's
> purposes--seeing the source code and testing its security is extremely
> important. "You wouldn't want to do it on Windows NT, because you know
> nothing about what is going on inside NT," he said.
>
> As if Ballmer didn't have enough headaches already... :)
>
> --
>
> Regards, Weird (Harold Stevens) * IMPORTANT EMAIL INFO FOLLOWS *
> Pardon the bogus email domain (dseg etc.) in place for spambots.
> Really it's (wyrd) at raytheon, dotted with com. DO NOT SPAM IT.
> Standard Disclaimer: These are my opinions not Raytheon Company.
Hello,
thx...I read the article and shudder:
"a more secure distribution
of Linux that the NSA has developed and an initial
version of which it released in December 2000. "
Wow, the only question is for whom is it much more secure and have they really
released it in december?
As I don't expect to find with a fast grep through the sources something like
this NSA_KEY.
Sure I could always diff those NSA sources against the once from kernel,org.
But I have no need as I personally trust those sources from Linus and all those
other great guys who make it happen,
much more than anything else, anyway was a nice read...:-)
Nice Weekend
Michael Heiming
--
When I die, I want to die peacefully, in my sleep. Like my grandfather.
Not screaming in terror, like the passengers in his car.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Convert Word-DOC to PostScript
Date: Sat, 03 Feb 2001 14:52:02 GMT
I'm lookin for a non-interactive process to do that
finding it tough to get it
In article <[EMAIL PROTECTED]>,
John Hasler <[EMAIL PROTECTED]> wrote:
> Bob Hauck writes:
> > The easy way is to load the file into StarOffice, select a PS
printer,
> > and print to a file.
>
> You left out the four days of continuous downloading to get
StarOffice.
>
> > You can even do this in Windows with Word if you have a PS printer
> > installed.
>
> If you have Windows installed.
> --
> John Hasler
> [EMAIL PROTECTED] (John Hasler)
> Dancing Horse Hill
> Elmwood, WI
>
Sent via Deja.com
http://www.deja.com/
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: CIFS and NFS file locking
Date: Sat, 3 Feb 2001 15:46:53 +0100
In comp.os.linux.misc Pawel Lewicki <[EMAIL PROTECTED]> wrote:
> Does anybody know any source of information about aspects of file locking in
> mentioned protocols?
Uh .. I have a big book on NFS in front of me. NFS Illustrated by
Brent Calaghan. Addison-Wesley Professional COmputing Series, December
1999.
And I had a huge book on samba too ...
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: User write privileges to /dos partitions
Date: Sat, 3 Feb 2001 15:48:46 +0100
Gregory Allen <[EMAIL PROTECTED]> wrote:
> I've browsed the how-to's and the FAQ's, trying different flags,
Why? This IS a faq.
> but to no avail.
> As root, I can read/write to /dos*. As joe user however, I can
> only read.
Of course.
> Can anyone suggest a fix so that "joe user" can also write to
> the /dos* partitions?
There is no fix. Just mount the partition with a umask that allows him
too! Man mount. Or mount it with a gid to which all users who should write
to the partition can belong to.
> I'm using slackware 7, if that matters.
> Thanks
> /dev/sdb4 /zip vfat user,noauto 0 1
> /dev/hda1 /dosc vfat defaults 1 0
> /dev/hda5 /dosd vfat defaults 1 0
Well, those are your problem!
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Modem [linmodem] configuration
Date: Sat, 03 Feb 2001 15:05:59 GMT
Panagiotis <[EMAIL PROTECTED]> wrote:
> HPATH = ../../include
> FINDHPATH = $(HPATH)/asm $(HPATH)/linux $(HPATH)/scsi $(HPATH)/net
This is a foulup.
> CC =$(CROSS_COMPILE)gcc -D__KERNEL__ -DMODULE -Wall $(HOSTCFLAGS) -I$(HPATH)
So is that.
> according to the manual the output shoould be a file in /root/lib called
> pctel.o which must
ignore them (and it won't be, unless you are in /root !). Just execute the
line
gcc -I/usr/src/linux-2.2.16/include -DMODULE -D__KERNEL__ -O2 -c foo.c
replacing the -I/... stuff with wherever the sources of your 2.2.16
kernel are.
> Where can I declare in the makefile the fact that my kernel is 2.2.16?
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: unapcking the 2.4 kernel in a home dir
Date: Sat, 3 Feb 2001 15:51:06 +0100
Paul Kimoto <[EMAIL PROTECTED]> wrote:
> In article <[EMAIL PROTECTED]>, jafgon wrote:
> README says:
> : Do NOT use the /usr/src/linux area! This area has a (usually
> : incomplete) set of kernel headers that are used by the library header
And this is a recent Redhat foulup, insanely immitated by debian, and
battily blessed by Linus. Go ahead and make /usr/src/linux a symlink to
yoru kernel area, like normal.
(yes, I see the reasoning, and no, I don't hold with it ... it worked
up till now, so why should we change it - it wrecks all third party
modules).
Peter
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: RAM advice
Date: Sat, 03 Feb 2001 10:15:44 -0500
Xavier Houppertz wrote:
>
> I still have 114 Mb but this is swapfile and thus much slower (but maybe
> linux handle this so good ...)
> I see that you have still 77 MB of free RAM, that's 'easy' !
> My memory usage doesn't change too much so I'am always 'on the edge' and
> when a user make a big query on the db, my server performance falls !
>
> About kswapd, it is running (don't know what it is anyway). I only attached
> the first page of the processes ...
>
> Xavier.
>
12:29pm up 72 days, 22:19, 11 users, load average: 0.25, 0.22, 0.16
109 processes: 96 sleeping, 2 running, 11 zombie, 0 stopped
CPU states: 11.5% user, 1.3% system, 0.0% nice, 87.4% idle
Mem: 257088K av, 252764K used, 4324K free, 95864K shrd, 14028K buff
Swap: 130748K av, 8184K used, 122564K free 100220K
cached
You do not understand. You are using only about 8 megabytes of your swap
file. the 100 Megabytes labelled as "cached" is main memory (RAM), NOT
swap space, presently being used for buffers and cache. These will be
released automatically by Linux if the space is required for something
more important.
Have you measured why your performance falls when making "a big query on
the db"? What dbms are you using? What is the nature of the query? Is it
compute-limited? Memory-limited? IO-limited?
I run IBM's DB2 UDB dbms that can beat the dickens out of this machine.
I can run a job that takes about 1/2 hour, running both 550MHz Pentium
III CPUs on this machine at about 99+% busy, and still do no paging. DB2
has about 12 processes running (not including the client), and even
though it does a lot of IO, it is still compute limited. With the amount
of memory I have in this machine, and due to DB2's good query
optimization, I would estimate that most of the data DB2 needs is kept
in the cache (not L1 or L2: the Linux cache). The disk IO rate would be
around 200K bytes/second on my 10,000 rpm Ultra-2 SCSI controllers, that
I have seen approach 20 Megabytes/second on IO-limited jobs. In theory,
my hard drives are capable of 40 Megabytes/second sustained transfer
rate, but I have never seen that.
My 72 megabytes of free memory were totally unused when I posted my
reply. I have more memory in my machine than I need, but it was so cheap
that it did not make sense to buy less. This way, if I "upgrade" to a
hungrier window manager than GNOME/Enlightenment, or start processing a
lot of moving video images, I will still have enough memory. And if not,
I have two memory slots open so I can go up to a Gigabyte of main
memory.
And I remember my first hard drive was 40 Megabytes of 2500 rpm stuff,
on a machine with 98,304 (6-bit) BYTES of RAM. That is BYTES, not
Kilobytes or Megabytes!
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 9:50am up 5 days, 18:19, 3 users, load average: 2.16, 2.11, 2.09
------------------------------
From: "Panagiotis" <[EMAIL PROTECTED]>
Subject: Re: Modem [linmodem] configuration
Date: Sat, 3 Feb 2001 17:12:55 +0200
so you mean i should just excecute
gcc -I/usr/src/linux-2.2.16/include -DMODULE -D__KERNEL__ -O2 -c foo.c
??
dont i have to edit anything?
------------------------------
From: [EMAIL PROTECTED] (Lee Webb)
Subject: Re: User write privilege on /dos partition
Date: 3 Feb 2001 15:19:14 GMT
Reply-To: [EMAIL PROTECTED]
On Fri, 02 Feb 2001 07:11:08 GMT, [EMAIL PROTECTED] wrote:
>I've browsed the how-to's and the FAQ's, trying different flags,
>but to no avail.
>
>As root, I can read/write to /dos*. As joe user however, I can
>only read.
>
>Can anyone suggest a fix so that "joe user" can also write to the
>/dos* partitions?
>
>I'm using slackware 7, if that matters.
>
>Thanks
>
>Here's an extract from my fstab:
>/dev/sda1 swap swap defaults 0 0
>/dev/sda2 / ext2 defaults 1 1
>/dev/sdb4 /zip vfat user,noauto 0 1
>/dev/hda1 /dosc vfat defaults 1 0
>/dev/hda5 /dosd vfat defaults 1 0
>/dev/hdc /cdrom iso9660 ro,user,noauto 0 1
>[snip]
>
/dev/hda1 /dosc vfat user 0 0 0
works fine for me.
Normal user can mount and read/write to /dosc.
Lee.
------------------------------
From: Eric Ho <[EMAIL PROTECTED]>
Subject: Re: Calendar Software
Date: 3 Feb 2001 15:22:48 GMT
Go to www.freshmeat.net and search for calandar.
Best Regards,
Eric Ho
[EMAIL PROTECTED] wrote:
> I am trying to find server and client software that allows for
> scheduling meetings,etc like the way MS Exchange and Outlook.
> Naturally I am looking for open-source based stuff. I would like to
> setup a calendar system in my home network so that my family can
> coordinate their schedules. I have a Linux server and different
> flavors of MS Windows clients (Windows 95, Windows 2000, Windows ME).
> Can anyone give me some pointers?
> Ron
> Sent via Deja.com
> http://www.deja.com/
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: unapcking the 2.4 kernel in a home dir
Date: Sat, 03 Feb 2001 10:25:46 -0500
"Peter T. Breuer" wrote:
>
> Paul Kimoto <[EMAIL PROTECTED]> wrote:
> > In article <[EMAIL PROTECTED]>, jafgon wrote:
> > README says:
>
> > : Do NOT use the /usr/src/linux area! This area has a (usually
> > : incomplete) set of kernel headers that are used by the library header
>
> And this is a recent Redhat foulup, insanely immitated by debian, and
> battily blessed by Linus. Go ahead and make /usr/src/linux a symlink to
> yoru kernel area, like normal.
>
My /usr/src/linux is a symlink to /usr/src/linux-2.2.14 (it is not as
bad as it looks, since VA Linux has put in most of the security updates
of later kernels, as well as other updates) that seems to be the
complete source tree for that version of the Linux kernel. Anyway, 71788
blocks of stuff. I imagine it would be larger if I had actually compiled
the kernel in there. This is VA Linux Systems version of RHL 6.2. VAL
calls it RHL 6.2.3.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 10:15am up 5 days, 18:44, 3 users, load average: 2.03, 2.20, 2.18
------------------------------
From: "Xavier Houppertz" <[EMAIL PROTECTED]>
Subject: Re: RAM advice
Date: Sat, 3 Feb 2001 16:27:19 +0100
> You do not understand. You are using only about 8 megabytes of your swap
> file. the 100 Megabytes labelled as "cached" is main memory (RAM), NOT
> swap space, presently being used for buffers and cache. These will be
> released automatically by Linux if the space is required for something
> more important.
Ok, got it !
> Have you measured why your performance falls when making "a big query on
> the db"? What dbms are you using? What is the nature of the query? Is it
> compute-limited? Memory-limited? IO-limited?
It is a sybase sql server (10.5).
the query is a select on several tables with compute and multiple order
clauses.
Nothing is limited so one query can eat 99.9% of the cpu time and memory.
I have no idea how to limit these (has to be in sybase central but where
...)
Maybe it would be smart to restrict the cpu and mem usage to 70% per
process.
Any comments ?
> I run IBM's DB2 UDB dbms that can beat the dickens out of this machine.
> I can run a job that takes about 1/2 hour, running both 550MHz Pentium
> III CPUs on this machine at about 99+% busy, and still do no paging. DB2
> has about 12 processes running (not including the client), and even
> though it does a lot of IO, it is still compute limited. With the amount
> of memory I have in this machine, and due to DB2's good query
> optimization, I would estimate that most of the data DB2 needs is kept
> in the cache (not L1 or L2: the Linux cache). The disk IO rate would be
> around 200K bytes/second on my 10,000 rpm Ultra-2 SCSI controllers, that
> I have seen approach 20 Megabytes/second on IO-limited jobs. In theory,
> my hard drives are capable of 40 Megabytes/second sustained transfer
> rate, but I have never seen that.
That's when you know your job and 'optimize' your hard and soft ... not my
case !
> My 72 megabytes of free memory were totally unused when I posted my
> reply. I have more memory in my machine than I need, but it was so cheap
> that it did not make sense to buy less. This way, if I "upgrade" to a
> hungrier window manager than GNOME/Enlightenment, or start processing a
> lot of moving video images, I will still have enough memory. And if not,
> I have two memory slots open so I can go up to a Gigabyte of main
> memory.
Think back when the RAM was about $50 the Mb ...
> And I remember my first hard drive was 40 Megabytes of 2500 rpm stuff,
> on a machine with 98,304 (6-bit) BYTES of RAM. That is BYTES, not
> Kilobytes or Megabytes!
Mine was only 20 Megabytes (hell of a speed compared to floppies)
Xavier.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.misc.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Misc Digest
******************************
