Linux-Development-Sys Digest #456, Volume #6 Sat, 6 Mar 99 01:14:12 EST
Contents:
scsi tape driver I/O error on mag tape ([EMAIL PROTECTED])
Re: lilo.conf (Shark)
Re: NFS Data Corruption between Linux 2.2.x and SunOS 5.5.1 (Tim Dawson)
ftp problem with redhat 5.2/mandrake (BL)
Re: ftp problem with redhat 5.2/mandrake (BL)
Re: booting NT & linux (Peter Samuelson)
Re: glibc dangling pointer? (Nate Eldredge)
Re: ftp problem with redhat 5.2/mandrake (jedi)
Bitches that look fucked up 97033 ([EMAIL PROTECTED])
Ethernet crc and frame errors? ("john")
Re: Any interest in a new POP3 server (Leslie Mikesell)
Re: Any interest in a new POP3 server (Peter Samuelson)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: scsi tape driver I/O error on mag tape
Date: Thu, 04 Mar 1999 22:17:33 GMT
Hi there,
I have a Digital TLZ07-DA tape drive attached to a Linux system (RH5.2). The
scsi tape driver is /lib/modules/2.0.36-0.7/scsi/st.o. Running the following
test program:
#include <sys/mtio.h>
#include <errno.h>
#define BUFSIZ 2048
#define ERPRINT printf("%s\n",strerror(errno))
int main(int argc, char **argv)
{
int fd;
int i;
char buf[BUFSIZ];
char * buf1 = "0009hello0006\r\n";
struct mtop op;
if ((fd = open("/dev/st0", 2, 511)) < 0) {
printf("unable to open: %d\n",fd);
ERPRINT;
exit(1);
}
/* write end-of-file record */
op.mt_op = MTWEOF;
op.mt_count = 1;
if ((i = ioctl(fd, MTIOCTOP, (char*)&op)) < 0) {
printf("unable to write EOF record: %d\n",i);
ERPRINT;
exit(1);
}
strcpy(buf, buf1);
if ((i = write(fd, buf, BUFSIZ)) != BUFSIZ) {
printf("error in writing: %d\n",i);
ERPRINT;
exit(1);
}
/* write forward space record */
op.mt_op = MTFSR;
op.mt_count = 1;
if ((i = ioctl(fd, MTIOCTOP, (char*)&op)) < 0) {
printf("unable to forward space: %d\n",i);
ERPRINT;
exit(1);
}
close(fd);
return 0;
}
I got the error:
unable to forward space: -1
Input/output error
I have two questions. Does the above program look OK in terms of the order of
operations? If yes, then is it a bug in st.o?
By the way, if I close the device without forwarding space in the above
program, then I can read what I have written to the tape. This seems to
vindicate the test program. But I am not an expert on mag tape...
Let me know if you have any thoughts on this issue. Thanks in advance!
Huayong
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: Shark <[EMAIL PROTECTED]>
Subject: Re: lilo.conf
Date: Fri, 05 Mar 1999 15:35:03 -0600
Reply-To: [EMAIL PROTECTED]
Did you run lilo after changing the file?
It should have said.
adding linux*
adding dos
Good luck
[EMAIL PROTECTED] wrote:
> Hello folks,
>
> I need help with configuring my lilo.conf file. The scenario is as follows:
>
> I had one SCSI disc dedicated to Linux. Recently I added a new disk to the
> system with DOS on it. Now I want to be able to go to DOS or Linux frm the
> Lilo prompt. I need help in setting up the lilo.conf script for this.
>
> My linux disk has SCSI ID 0
> My DOS disk has SCSI ID 1
>
> The DOS disk can be mounted as sdb1, but I can't seem to get the script right
> for accessing DOS from lilo prompt.
>
> The lilo.conf on my system is shown below:
>
> boot=/dev/sda
> map=/boot/map
> install=/boot/boot.b
> prompt
> timeout=50
> image=/boot/vmlinuz-2.0.34-0.6
> label=linux
> root=/dev/sda1
> initrd=/boot/initrd-2.0.34-0.6.img
> read-only
> other=/dev/sdb1
> label=dos
> table=/dev/sdb
>
> SOS
> Manoj Panicker
>
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED] (Tim Dawson)
Subject: Re: NFS Data Corruption between Linux 2.2.x and SunOS 5.5.1
Date: 6 Mar 1999 01:49:10 GMT
Based on several responses to my original replay to this thread, I will say it
again. This is NOT just a Solaris bug! There are at least 6 persons other than
myself who have seen this problem with Linux <--> Linux NFS with the 2.2.x client.
(Server code does not seem to matter.) Hence, while Sun may have a patch to
work around this issue, I feel that there is a much deeper problem going on here.
With just 2.2.x, my 'discovery' case was a wine compile. With this kernel, NFS
would hang solid on many files.
With 2.2.x-ac5, the hangs went away, but instead files were truncated and the
compile corrupted.
With the 2.0.36 kernel/client, everything build perfectly. Kernels were build
from the same (as near as possible) and tests on the same system. Server was
Linux 2.0.36 or 2.2.x with either the kernel or user space NFS daemons. Server
variations did not affect the problem.
One of the other sites that I know to have this problem is Alexa Internet, in
San Francisco. If anyone in the kernel group would like more info on our
experiences/tests, please let me know and I will forward contact info ASAP, as
well as more detail on my testing.
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Michal Szymanski) writes:
|> On Thu, Feb 18, 1999 at 07:27:35PM +0000, Linus Torvalds wrote:
|> > In article <[EMAIL PROTECTED]>,
|> > Oliver Stahlhut <[EMAIL PROTECTED]> wrote:
|> > >
|> > >Data written to a SUN Solaris NFS server by a Linux 2.2.1 NFS client can
|> > >be truncated.
|> >
|> > It's a Solaris bug. See the attached information about official Sun bug
|> > ID numbers etc, and just contact your Sun representative.
|>
|> Two questions:
|>
|> 1. Why older Linux kernels (2.0.XX) worked fine, if it is just a
|> Solaris bug? I used to compile big packages on a Linux machine
|> RH 4.2/2.0.35 on NFS-mounted directory on a Solaris server
|> and it worked rock-solid. Now with 2.2.1 I can just forget
|> about doing such things.
|>
|> 2. Any luck in finding the patch for Solaris 2.5.1 SPARC?
|>
|> One comment:
|>
|> For years, Linux NFS server was unable to serve Solaris clients properly
|> (a topic discussed many times on Usenet - any attempt to write
|> a hundred-MB file ended up in "linux server not responding")
|> Finally, with 2.2.1 and knfsd it started to work as expected.
|> And now, it is broken the other way. ;-(((((((((((((((((((((
|>
|> regards,
|>
|> Michal.
|>
|> --
|> Michal Szymanski ([EMAIL PROTECTED])
|> Warsaw University Observatory, Warszawa, POLAND
--
================================================================================
Tim Dawson ([EMAIL PROTECTED]) Owner/Engineer
TPC Services Bellnet: (972)-221-7385
Lewisville, Texas 75067 FAXnet: (972)-221-0393
"The world is complex. Sendmail.cf reflects this...."
------------------------------
From: BL <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: ftp problem with redhat 5.2/mandrake
Date: 6 Mar 1999 03:06:21 GMT
after a new install of 5.2 RH (I used 4.2 for the last 2 yrs) my inbound ftp
is not working. no matter what I try, the connection is refused.
what things should I look for? is PAM more picky than it was in RH4.2 days?
I'm not even trying to ftp to the root acct, I created a nonpriv user and ftpd
won't even let me login as that..
in.ftpd is there and inetd.conf has the entry uncommented. the user is NOT in
/etc/ftpusers. ftphosts is empty. ftpgroups is empty.
what else does redhat5.2 (mandrake, actually) need so that I can enable
inbound ftp's to my site?
(is there a bug in the standard shipped in.ftp on redhat 5.2, maybe?)
thanks,
--
.bl
------------------------------
From: BL <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: ftp problem with redhat 5.2/mandrake
Date: 6 Mar 1999 04:19:26 GMT
In comp.os.linux.networking jedi <[EMAIL PROTECTED]> wrote:
: On 6 Mar 1999 03:06:21 GMT, BL <[EMAIL PROTECTED]> wrote:
: >after a new install of 5.2 RH (I used 4.2 for the last 2 yrs) my inbound ftp
: >is not working. no matter what I try, the connection is refused.
: >
: >what things should I look for? is PAM more picky than it was in RH4.2 days?
: >
: >I'm not even trying to ftp to the root acct, I created a nonpriv user and ftpd
: >won't even let me login as that..
: >
: >in.ftpd is there and inetd.conf has the entry uncommented. the user is NOT in
: >/etc/ftpusers. ftphosts is empty. ftpgroups is empty.
: >
: >what else does redhat5.2 (mandrake, actually) need so that I can enable
: >inbound ftp's to my site?
: Make an appropriate subdirectory in /home/ftp/pub/.
I'll have to try that on another install. the way I fixed this problem was to
download the SRPM and rebuild from src. it then finally worked ;-)
so are you saying it was a CWD problem? strange... you'd think that the
standard redhat install would not have this problem. I can't believe I'm the
only one seeing ftpd failures out-of-the box.
thanks,
--
.bl
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Crossposted-To: comp.os.ms-windows.nt.admin.misc,comp.os.ms-windows.nt.misc
Subject: Re: booting NT & linux
Date: 5 Mar 1999 22:41:49 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[<[EMAIL PROTECTED]>]
> NT's boot manager will work fine. What you need to do is set Linux
> to boot from the SCSI disk boot sector and create a bootsector file
> for NT's boot loader.
Assuming the SCSI adapter installs a BIOS so the drive is usable at
boot time, of course.
> First you need to disable the IDE HD in the BIOS (or better yet set
> SCSI to boot) and boot Linux. Then you run LILO to create a boot
> sector on your SCSI disk (read the docs if you need to). You then
> use dd to create a disk file of the boot sector. Substitute your
> SCSI drive device name where appropriate. dd if=/dev/sda
> of=/bootsect.lnx bs=512 count=1
Or ... tell LILO to write the boot sector in the right place so you
don't have to mess with copying it. This has the advantage that things
don't break when you rerun LILO.
Basically, make sure your FAT partition is mounted (probably might as
well have it in /etc/fstab). If drive C is NTFS ignore this message --
NTFS R/W support was very buggy last I checked. Create the file
/mnt/c/bootsect.lnk (or whatever) with
dd bs=512 count=1 if=/dev/zero of=/mnt/c/bootsect.lnx
and then start /etc/lilo.conf with the line
boot=/mnt/c/bootsect.lnx
Then of course edit /mnt/c/boot.ini as already mentioned:
[boot loader]
timeout=3
default=c:\bootsect.lnx
[operating systems]
c:\bootsect.lnx="Linux"
multi(0)disk(0)rdisk(0)partition(1)\WINNT="Legacy Apps"
It may be possible even with an NTFS partition to get the NT loader to
load LILO from your SCSI disk bootsector -- play with multi(X)disk(X)
settings to see if it sees any of them as the right one.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
Date: Thu, 04 Mar 1999 16:13:10 -0800
From: Nate Eldredge <[EMAIL PROTECTED]>
Subject: Re: glibc dangling pointer?
Greg Herlein wrote:
>
> I am having a problem: I get a core dump if I link my app to ccmalloc (a
> leak checker). ccmalloc has worked on my older Linux boxes, and it works
> fine on Solaris, but on either of two Red Hat 5.2 boxes (stock) it fails
> the same way. It almost smells like there is a glibc issue here, but I'm
> not smart enough about that kind of thing to know - hence this post.
>
> Here is the gdb stack trace:
>
> Program terminated with signal 11, Segmentation fault.
> Reading symbols from /lib/libdl.so.2...done.
> Reading symbols from /lib/libc.so.6...done.
> Reading symbols from /lib/ld-linux.so.2...done.
> #0 0x4002e811 in __kill ()
> (gdb) bt
> #0 0x4002e811 in __kill ()
> #1 0x809c246 in die () at callchain.c:517
> #2 0x80a056d in _ccmalloc_free (_data=0x832ec80) at callchain.c:4676
> #3 0x80a074e in ccmalloc_free (_data=0x832ec80) at callchain.c:4847
> #4 0x809c098 in free (p=0x832ec80) at wrapper.c:391
> #5 0x4006af71 in __tzfile_read (file=0x400a2194 "/etc/localtime")
> at tzfile.c:97
> #6 0x4006a083 in tzset_internal (always=1) at tzset.c:204
> #7 0x4006ae1c in __tz_convert (timer=0xbffff93c, use_localtime=1,
> tp=0x400a6910) at tzset.c:617
> #8 0x40067742 in localtime (t=0xbffff93c) at localtime.c:48
> #9 0x40067619 in ctime (t=0xbffff93c) at ctime.c:32
> #10 0x804a0bc in PrintIntroduction (pf=0x80f67d8) at iserver.c:368
> #11 0x8049d8b in main (argc=1, argv=0xbffffa70) at iserver.c:185
>
> Does this make sense at all? Why die in kill()???
Probably because it raises SIGSEGV to kill itself and dump core. If I
were writing ccmalloc, I would have inserted a message at that point to
give the user some idea what is wrong.
> I might add that if I
> don't link ccmalloc to it, it runs fine. I just want to chase down any
> memory leaks!
>
> I have reported this to the author of ccmalloc (Armin Biere
> <[EMAIL PROTECTED]>) and he said that he suspected glibc. I would tend to
> agree since it does not happen on a stock SOlaris machine.
What glibc does RH 5.2 have?
A bug in glibc is, of course, possible, but it's also possible that it
could be a bug in ccmalloc. Possibly one that doesn't show up under
other systems.
If you have the ccmalloc source, you could poke around with your
debugger and try to find out what's wrong.
Also, glibc has features built in to it that can allow to check memory
leaks and more. It's all in the manual.
--
Nate Eldredge
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (jedi)
Crossposted-To: comp.os.linux.networking
Subject: Re: ftp problem with redhat 5.2/mandrake
Date: Fri, 5 Mar 1999 20:37:59 -0800
On 6 Mar 1999 04:19:26 GMT, BL <[EMAIL PROTECTED]> wrote:
>In comp.os.linux.networking jedi <[EMAIL PROTECTED]> wrote:
>: On 6 Mar 1999 03:06:21 GMT, BL <[EMAIL PROTECTED]> wrote:
>: >after a new install of 5.2 RH (I used 4.2 for the last 2 yrs) my inbound ftp
>: >is not working. no matter what I try, the connection is refused.
>: >
>: >what things should I look for? is PAM more picky than it was in RH4.2 days?
>: >
>: >I'm not even trying to ftp to the root acct, I created a nonpriv user and ftpd
>: >won't even let me login as that..
>: >
>: >in.ftpd is there and inetd.conf has the entry uncommented. the user is NOT in
>: >/etc/ftpusers. ftphosts is empty. ftpgroups is empty.
>: >
>: >what else does redhat5.2 (mandrake, actually) need so that I can enable
>: >inbound ftp's to my site?
>
>: Make an appropriate subdirectory in /home/ftp/pub/.
>
>I'll have to try that on another install. the way I fixed this problem was to
>download the SRPM and rebuild from src. it then finally worked ;-)
>
>so are you saying it was a CWD problem? strange... you'd think that the
>standard redhat install would not have this problem. I can't believe I'm the
>only one seeing ftpd failures out-of-the box.
On redhat 5.x, anon ftp should be enabled and active by default.
Uploads merely require a place for them to go where the guest
user has write priveledge.
Mandrake may have changed this for security purposes.
--
Herding Humans ~ Herding Cats
Neither will do a thing unless they really want to, or |||
is coerced to the point where it will scratch your eyes out / | \
as soon as your grip slips.
In search of sane PPP docs? Try http://penguin.lvcm.com
------------------------------
From: [EMAIL PROTECTED]
Subject: Bitches that look fucked up 97033
Date: Thu, 04 Mar 1999 17:46:19 -0600
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.development.apps
Passwords are kick'n
http://sexshare.com/~stmonkey/password.htm
Top 10 Sex Sites exposed for your Enjoyment--==>
http://sexshare.com/~stmonkey/password.htm
2,h&QR'6K"
------------------------------
From: "john" <[EMAIL PROTECTED]>
Subject: Ethernet crc and frame errors?
Date: Thu, 4 Mar 1999 20:04:15 -0500
I'm writing a protocol layer that uses an in-house messaging format.
Eventually, this layer will run over a high speed switching core, but, for
now, I'm using ethernet, (until the real hardware becomes available).
Anyways, I've ported LAM61 (MPI) to use this new protocol, but occasionally
I'll get a spat of 4 or 5 ethernet crc/frame errors, after sending a series
of packets.
I'm using the 3c59x device driver.
Can anyone help me????
John
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Subject: Re: Any interest in a new POP3 server
Date: 5 Mar 1999 23:50:49 -0600
In article <[EMAIL PROTECTED]>,
Jim Abbey <[EMAIL PROTECTED]> wrote:
>> > > It needs to also support qmail maildirs.
>> For use with qmail? Not everyone uses Sendmail, and there are many
>> good reasons for that. Qmail has some nice features.
>
>Not everyone. But is it reasonable to say "most people?"
Qmail has some odd concepts, but the idea of delivering one
message per file eliminates locking problems and should work
reliably over NFS.
>> It is a separate issue whether the world needs another IMAP server or,
>> for that matter, another POP3 server. Or whether the two servers would
>> share enough code to justify being one executable. (Sendmail taught us
>> the dangers of having one binary try to do too many things.) I think a
>> combined server would be worthwhile, personally.
>>
>
>Sendmail is incomprehensible. IMAP is, as you imply, headed firmly in
>the same direction. Could POP3 be the VHS to IMAP's Betamax? I've always
>believed in KISS.
POP is a reasonable access protocol if you always want to copy
every message over to the client machine. IMAP is designed for
people who use more than one computer (doesn't everybody?) and
want consistent views of the same mailbox(es) from everywhere.
The messages can remain on the server where presumably they
are easier to keep backed up for everyone. If you sometimes
work over low bandwidth links you can avoid downloading the
large attachments to those locations unless you want them.
By the way, how would this server be different than the UW and
Cyrus servers?
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Any interest in a new POP3 server
Date: 4 Mar 1999 18:33:57 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Phil Howard]
> > It needs to also support qmail maildirs.
[Jim Abbey <[EMAIL PROTECTED]>]
> Why's that?
For use with qmail? Not everyone uses Sendmail, and there are many
good reasons for that. Qmail has some nice features.
> > How about make it an integrated POP+IMAP server?
> Is IMAP in widespread use?
Widespread enough -- I know of at least three clients that can use it:
Pine, Netscape Mail and Emacs VM. Besides, if it is indeed a superior
protocol (which was the goal of its designers, but I haven't studied
POP3 enough to really say for sure), that in itself justifies
supporting it -- build a server and the clients will come.
It is a separate issue whether the world needs another IMAP server or,
for that matter, another POP3 server. Or whether the two servers would
share enough code to justify being one executable. (Sendmail taught us
the dangers of having one binary try to do too many things.) I think a
combined server would be worthwhile, personally.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
** 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 (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
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-Development-System Digest
******************************
