Linux-Networking Digest #177, Volume #12 Tue, 10 Aug 99 16:13:53 EDT
Contents:
Re: How to Log attemped connections (Jose Nazario)
Re: cbq, bandwithshaping possible bugs in kernel (Leon Harris)
using setserial to change irq on serial port (Jan Cernohorsky)
Re: Feasibility for ip-masq Fire (QuestionExchange)
Re: HELP with firewalls ("ICON Business Internet Services Engineering")
Re: Help - Stupid question(s) (QuestionExchange)
Re: getting network device info (QuestionExchange)
Re: Installing an ethernet card (QuestionExchange)
Re: Linux as NFS-Server for IRIX (QuestionExchange)
Re: Linux and MS Proxy (Intellectronix)
Re: IP Masq and ICQ Chat (QuestionExchange)
Re: Linux Printing to a Remote Printer ("F.B.Quinn")
info about devpts filesystem desired (Mark Johnson)
Re: Q (QuestionExchange)
Re: IP-Number for device (QuestionExchange)
DHCP and RedHat 6.0, no gateway or DNS assigned??? (bizarrod)
----------------------------------------------------------------------------
From: Jose Nazario <[EMAIL PROTECTED]>
Crossposted-To: comp.security.firewalls,comp.security.unix
Subject: Re: How to Log attemped connections
Date: Tue, 10 Aug 1999 13:57:12 -0400
Matt wrote:
> Is there a way to log all attemped connections and what ports they were on
> under Linux?
use xinetd which has enhanced logging as one of it's fantastic
enhancements over inetd. read it's logs... works extremely well.
for the pure fascist and disk space wasting types, IP accounting.
jose nazario [EMAIL PROTECTED]
------------------------------
Date: Tue, 10 Aug 1999 11:36:52 +0800
From: Leon Harris <[EMAIL PROTECTED]>
Subject: Re: cbq, bandwithshaping possible bugs in kernel
Ok, got it sussed.
I guess I had better answer this for the next poor sod who gets as badly
tied up as I was.
It turns out that there are differences in how gcc and egcs optimize
code - the egcs apparently fails to optimize the cbq code in the kernel
properly. Hence this obscure and nasty bug.
Two solutions are apparently possible: use gcc 2.7.2 to compile kernels
with class based queuing in it.
Or, apply the following patch ( Much thanks to Alexey Kuznetsov for his
patch, and kind help). ( you may need to fiddle with the paths on the
diffs to make it take - I had to)
diff -ur ../orig/linux/include/asm-i386/system.h
linux/include/asm-i386/system.h
--- ../orig/linux/include/asm-i386/system.h Wed May 12 22:40:52 1999
+++ linux/include/asm-i386/system.h Wed Jul 14 16:41:01 1999
@@ -129,24 +129,26 @@
/*
* Note: no "lock" prefix even on SMP: xchg always implies lock anyway
+ * Note 2: xchg has side effect, so that attribute volatile is
necessary,
+ * but generally the primitive is invalid, *ptr is output
argument. --ANK
*/
static inline unsigned long __xchg(unsigned long x, void * ptr, int
size)
{
switch (size) {
case 1:
- __asm__("xchgb %b0,%1"
+ __asm__ __volatile__("xchgb %b0,%1"
:"=q" (x)
:"m" (*__xg(ptr)), "0" (x)
:"memory");
break;
case 2:
- __asm__("xchgw %w0,%1"
+ __asm__ __volatile__("xchgw %w0,%1"
:"=r" (x)
:"m" (*__xg(ptr)), "0" (x)
:"memory");
break;
case 4:
- __asm__("xchgl %0,%1"
+ __asm__ __volatile__("xchgl %0,%1"
:"=r" (x)
:"m" (*__xg(ptr)), "0" (x)
:"memory");
diff -ur ../orig/linux/include/net/pkt_cls.h linux/include/net/pkt_cls.h
--- ../orig/linux/include/net/pkt_cls.h Mon Apr 5 19:34:14 1999
+++ linux/include/net/pkt_cls.h Wed Jul 14 16:39:12 1999
@@ -79,9 +79,12 @@
extern __inline__ unsigned long cls_set_class(unsigned long *clp,
unsigned long cl)
{
- cl = xchg(clp, cl);
+ unsigned long old_cl;
+
+ old_cl = *clp;
+ *clp = cl;
synchronize_bh();
- return cl;
+ return old_cl;
}
extern int register_tcf_proto_ops(struct tcf_proto_ops *ops);
diff -ur ../orig/linux/net/sched/sch_cbq.c linux/net/sched/sch_cbq.c
--- ../orig/linux/net/sched/sch_cbq.c Sat May 1 16:38:30 1999
+++ linux/net/sched/sch_cbq.c Sat Jul 17 20:53:17 1999
@@ -176,6 +176,7 @@
struct cbq_class *tx_borrowed;
int tx_len;
psched_time_t now; /* Cached timestamp */
+ psched_time_t now_rt; /* Cached real time */
unsigned pmask;
struct timer_list delay_timer;
@@ -375,9 +376,11 @@
if (toplevel > cl->level && !(cl->q->flags&TCQ_F_THROTTLED)) {
psched_time_t now;
+ psched_tdiff_t incr;
+
PSCHED_GET_TIME(now);
- if (PSCHED_TLESS(now, q->now))
- now = q->now;
+ incr = PSCHED_TDIFF(now, q->now_rt);
+ PSCHED_TADD2(q->now, incr, now);
do {
if (PSCHED_TLESS(cl->undertime, now)) {
@@ -503,7 +506,7 @@
}
}
- q->wd_expires = delay;
+ q->wd_expires = base_delay;
}
}
@@ -756,14 +759,19 @@
idle = (now - last) - last_pktlen/rate
*/
- idle = PSCHED_TDIFF(q->now, cl->last) - L2T(cl, len);
+ idle = PSCHED_TDIFF(q->now, cl->last);
+ if ((unsigned long)idle > 128*1024*1024) {
+ avgidle = cl->maxidle;
+ } else {
+ idle -= L2T(cl, len);
/* true_avgidle := (1-W)*true_avgidle + W*idle,
where W=2^{-ewma_log}. But cl->avgidle is scaled:
cl->avgidle == true_avgidle/W,
hence:
*/
- avgidle += idle - (avgidle>>cl->ewma_log);
+ avgidle += idle - (avgidle>>cl->ewma_log);
+ }
if (avgidle <= 0) {
/* Overlimit or at-limit */
@@ -980,10 +988,13 @@
struct sk_buff *skb;
struct cbq_sched_data *q = (struct cbq_sched_data *)sch->data;
psched_time_t now;
+ psched_tdiff_t incr;
PSCHED_GET_TIME(now);
+ incr = PSCHED_TDIFF(now, q->now_rt);
if (q->tx_class) {
+ psched_tdiff_t incr2;
/* Time integrator. We calculate EOS time
by adding expected packet transmittion time.
If real time is greater, we warp artificial clock,
@@ -991,12 +1002,14 @@
cbq_time = max(real_time, work);
*/
- PSCHED_TADD(q->now, L2T(&q->link, q->tx_len));
- if (PSCHED_TLESS(q->now, now))
- q->now = now;
+ incr2 = L2T(&q->link, q->tx_len);
+ PSCHED_TADD(q->now, incr2);
cbq_update(q);
- } else if (PSCHED_TLESS(q->now, now))
- q->now = now;
+ if ((incr -= incr2) < 0)
+ incr = 0;
+ }
+ PSCHED_TADD(q->now, incr);
+ q->now_rt = now;
for (;;) {
q->wd_expires = 0;
@@ -1044,6 +1057,11 @@
del_timer(&q->wd_timer);
if (delay <= 0)
delay = 1;
+ if (delay > 10*HZ) {
+ if (net_ratelimit())
+ printk(KERN_DEBUG "CBQ delay %ld
> 10sec\n", delay);
+ delay = 10*HZ;
+ }
q->wd_timer.expires = jiffies + delay;
add_timer(&q->wd_timer);
sch->flags |= TCQ_F_THROTTLED;
@@ -1224,7 +1242,7 @@
struct cbq_class *cl, *cl_head;
int prio;
- for (prio = TC_CBQ_MAXPRIO; prio >= 0; prio++) {
+ for (prio = TC_CBQ_MAXPRIO; prio >= 0; prio--) {
if ((cl_head = q->active[prio]) == NULL)
continue;
@@ -1252,6 +1270,8 @@
del_timer(&q->wd_timer);
del_timer(&q->delay_timer);
q->toplevel = TC_CBQ_MAXLEVEL;
+ PSCHED_GET_TIME(q->now);
+ q->now_rt = q->now;
for (prio = 0; prio <= TC_CBQ_MAXPRIO; prio++)
q->active[prio] = NULL;
@@ -1425,6 +1445,8 @@
q->delay_timer.data = (unsigned long)sch;
q->delay_timer.function = cbq_undelay;
q->toplevel = TC_CBQ_MAXLEVEL;
+ PSCHED_GET_TIME(q->now);
+ q->now_rt = q->now;
cbq_link_class(&q->link);
------------------------------
From: Jan Cernohorsky <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.misc
Subject: using setserial to change irq on serial port
Date: Tue, 10 Aug 1999 18:30:43 GMT
I have a SuSE 6.2 Linux installation on a dual processor PII 300 box. The
modem is ISA and on COM3. In the standard setting the irq on the port in
question would like to be irq 4. It needs to be 5. I have tried to use
setserial /dev/ttyS2 irq 5
(as root), but it comes back with
"Operation not permitted"
Permissioning as such should not be a problem since I attempt this as
root. Even worse, if I try to call setserial at boot time through the
relevant script in /etc/rc.d/serial
I get the same negative result. Anybody any idea what may be causing this
goofy behaviour of setserial, or any other way to reset my IRQ on that
port ?
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: Feasibility for ip-masq Fire
Date: 10 Aug 1999 18:48:13 GMT
> hello, all, I have about 100 users who need internet access
> (mostly web browsing), I am thinking of pulling in a SDSL, let
> say 768k/768k, and setup a IP-Masquerade Firewall box used as
> gateway to the internet. Is this a feasible thing to do? What
> is the feasible hareware requirement for this Linux box. any
> suggestion is appreciated. Li
Sounds good to me. The internet pipe you plan should be able
to handle the load. I would use a dual CPU platform with RH
Linux and Apache as the web server. Remember to set the
..hhtp/httpd.conf file to spawn about 50 deamons to start and a
max cound of 100 to groe into. Your system is going to need at
least 128m of memory for all the users. There are many firewall
products that will fit this configuration. Memory is the key
item here. Good Luck... --miked348
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2432&cus_id=USENET&qtn_id=1577
------------------------------
From: "ICON Business Internet Services Engineering" <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.misc
Subject: Re: HELP with firewalls
Date: Tue, 10 Aug 1999 10:41:39 -0500
@home handles hosts with just the hostname because they expect you to set
your DNS domain to what they tell you (or they set it via DHCP). They have
many sub-domains setup, one or more for each city you are in. Your mail and
news servers are specific to your locale. You can usually fix this in
either of two ways. First you can use the Fully Qualified Domain Name
(FQDN) for these servers in the configuration of mail and news clients. You
do this by putting the local domain after the server name. like
news.dlls1.tx.home.com for a news server in Dallas TX, or
mail.plano1.tx.home.com for a mail server in Plano TX. Your second option
is to add the assigned domain from @home to your domain search list where
you setup DNS. On your linux box this should be /etc/resolv.conf. It
should look something like this:
domain mydomain.com
search mydomain.com MY_CITY1.MY_ST.HOME.COM
nameserver 198.6.1.1
nameserver 198.6.1.2
nameserver 198.6.1.2
for the nameserver lines, you should use those provided by @home, they will
be faster. If you domain line matches the domain given to you by @home, you
should not have this problem.
I hope this helps. It works for me..
thebrownhighlander <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> HI,
>
> I have this firewall up. I got a script from somewhere and am just
> modifying it. Anyway, I am ip masqing with @home using redhat 5.2. My
> firewall is ok for everything else except for the mail and news servers.
>
> On my win box, when I try to connect to nntp server as "news" ("news" is
> how @home describes their news server) it won't let me. But when I log on
> using the ip address of the nntp server it works fine. This is the
> following for the firewall script
>
> # NNTP NEWS client (119)
> # ----------------------
> ipfwadm -I -a accept -P tcp -k -W $EXTERNAL_INTERFACE \
> -S $NEWS_SERVER 119 \
> -D $IPADDR $UNPRIVPORTS
>
> ipfwadm -O -a accept -P tcp -W $EXTERNAL_INTERFACE \
> -S $IPADDR $UNPRIVPORTS \
> -D $NEWS_SERVER 119
>
> where NEWS_SERVER = "news"
> EXTERNAL_INTERFACE = "eth1"
> UNPRIVPORTS = "1024:65535"
>
> now I have tried changing NEWS_SERVER to its corresponding ip address but
> that did not seem to work. It is only when I explicitly say my nntp news
> server is "news"
>
> However on my linux box when I ping "news" it pings fine (this is how I
> got ip address)
>
> Likewise it is the same for my POP3_SERVER and SMTP_SERVER
>
> Anyway I created this firewall from this place:
> http://rlz.ne.mediaone.net/linux/firewall/
>
> Please respond to this query.
>
> Thank you
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: Help - Stupid question(s)
Date: 10 Aug 1999 18:48:29 GMT
> Ok, still learning . . . . and have couple problems after
> days, weeks, and little over a month of HOW-TOs and
> configuration. A) Can ping network and all workstations from
> anywhere, even using names vs numbered addresses. B) Have
> access to network 'home' directories from wksta. C) Can
> assign 'drive' on wksta to Linux /home/* files. D) 'Think"
> ipchains and MASQ working - but . . . 1) How do I use Netscape
> from the clients? a) Have set up proxy pointing to host,
> much activity on hub, but no DNS found. b) Do I need telnet
> access on the HOST? 2) Kppp does it's own file configuration
> on activation, such as resolv.conf. PROBLEM is about 30% of the
> time - a connection is made to my LOCAL net (class C), and the
> home address of the ISP ONLY!! Cannot locate (or ping)
> addresses outside of my localedomain or Pronetisp.net - (been
> chasing this one for 2 weeks). ISP tells me they cannot help,
> even tho they're running RedHat . Any Guru ideas ?? I'M
> ALMOST THERE ! ! Really want to cut WinDo$e lose <G>
Well, since your killing off windows, ill help ;-) I assume
since your using ipchains, linux is a gateway. If thats the
case, i would configure it as a DHCP server, and report the DNS
server of your ISP to all users. They (the clients) are
probably having problems with using your linux box as a
cacheing non relay server, which means, that which your linux
box doesnt know, they cant. tell them to bypass linux as a DNS
server and if that fixes it, check your configs to make sure
linux has your ISP as one of its DNS's (linuxconf)
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2440&cus_id=USENET&qtn_id=1582
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: getting network device info
Date: 10 Aug 1999 18:48:46 GMT
> I am trying to aquire details of all (or at least some)
> devices on a given network, including identifying gateways
> from "normal" hosts, identifying print servers and locating and
> getting router details. Is this information to be found in the
> standard networking files or am I looking in the wrong place?
> -- Alex Sent via Deja.com http://www.deja.com/ Share what you
> know. Learn what you don't.
Try the command netstat. It will return all the details you
look for. Try in this order: mysys% netstat -r # Routing
table mysys% netstat # All the stuff Goof Luck --
miked348
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2434&cus_id=USENET&qtn_id=1583
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: Installing an ethernet card
Date: 10 Aug 1999 18:49:6 GMT
> Greetings - I am a linux novice trying to install an ethernet
> card in a linux machine. The hardware is a "thelinuxstore.com"
> pre-install machine with Red Hat 6.0. The network card is an
> Addtron AEF-360TX (PCI). I've installed the card, and have a
> light on both the card and the hub. During Boot-Up it tries to
> initialize or recognize the card and reports "Failed". I've
> tinkered with ifconfig and some of the control panel-type
> utilities, but I'm getting nowhere. I'm blundering around in
> there like a bull in a china shop, and something may break
> soon! Is that card even supported? Should I be using a certain
> tool or procedure to initialize the interface? Do I need some
> type of driver specific to the card? Many thanks in advance!
> bill hoey b0mbshelter industrial technologies
> [EMAIL PROTECTED] -----------== Posted via
> Newsfeeds.Com, Uncensored Usenet News ==----------
> http://www.newsfeeds.com The Largest Usenet Servers in
> the World! ------== Over 73,000 Newsgroups - Including
> Dedicated Binaries Servers ==-----
Check the .etc/conf.modules file for entry for the actual card
and make sure yoy have a entry for ne2000. Also double check
with the following commands: ins mod ne2000.o Good Luck...
--miked348
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2439&cus_id=USENET&qtn_id=1586
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: Linux as NFS-Server for IRIX
Date: 10 Aug 1999 18:48:52 GMT
> Hi outthere, I don't know if someone has solved it jet, but
> it seems at least the to people I send this mssg to have
> similar problems as I. I use a linux box as an NFS server for
> IRIX, linux and OSF/1 -Machines. It got it running with the
> kernel-based NFS-Server in SuSE 6.1 (suse may not very popular,
> but I'm sure redhat has a kernel-NFS, too) The "permission
> denied" Peter Camenzind gets when mounting, may be due to using
> different nfs-versions. IRIX needs "nfs2" as filesystem type in
> fstab, it is "nfs vers=2" in OSF/1. Maybe something similar on
> Solaris. By the way Kernel-based NFS-Server needs a >2.2.X
> Kernel, so 2.0.36 won't work. My problem now is : The linux-
> nfs server drops down after a few days. When wotring on the
> IRIX-Machine I get "stale nfs handle; unable to determine
> current directory" smthing like that. Then I try to restart
> the server on the linux box, but it complains:
> /sbin/init.d/nfsserver restart Shuting down [...] cannot[...]:
> address already in use. after that, only a reboot helps me out
> (for about 2 days...) Any hints ? Thanks, cr -- Christian
> Rummey, Dipl. Chem. Institut fuer Organische Chemie,
> Universitaet Wuerzburg mailto:[EMAIL PROTECTED]
Sounds like your NFS server is getting locked up with users not
closing down. I would suggest a script using fuser command to
remove connections that have gone stale or disconnect all when
you at least restart. Add the fuser command with correct list
in start/stop/restart rc script. Check users to verify this
time up related issue. Good Luck... --miked348
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2436&cus_id=USENET&qtn_id=1584
------------------------------
From: Intellectronix <[EMAIL PROTECTED]>
Crossposted-To: alt.linux,alt.os.linux,microsoft.public.backoffice.smallbiz
Subject: Re: Linux and MS Proxy
Date: Tue, 10 Aug 1999 14:34:58 +0100
Reply-To: [EMAIL PROTECTED]
You could use JProxyma instead of MS Proxy.
Please check out http://www.intellectronix.com/jpro/
David Eno wrote:
> I'm running MS Small Business BackOffice Server 4.0. I'm running the
> version of MS Proxy that comes with SBBO. I have a few Linux boxes that I
> would like to get to the WWW from. It appears that MS Proxy is not letting
> Netscape on the Linux machines to get to the internet.
>
> Is there anything I can do to make Linux be able to get to the internet
> through MS Proxy?
>
> BTW, I have 20 Win98 boxes that work fine on the network.
>
> TIA for you help.
>
> --
> Dave E.
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: IP Masq and ICQ Chat
Date: 10 Aug 1999 18:49:19 GMT
> I am running a router with IP masquerading on my LAN. I know
> it doesnt work with ICQ filetransfer (downloading to me) but I
> am also having problems sometimes with the ICQ chat system
> especially if its with more than one person. It seems to
> timeout and so tells me people have gone off the chat when they
> have not. Other times when joining or starting a chat some
> people will NEVER show up yet other people on the chat can see
> them. If I cannot see someone they cannot see me either and it
> appears to be trouble with the way ICQ does a direct
> connection. Other times it works fine though and recently and
> had a chat with 7 other people (the most I ever I have done)
> and it worked perfectly. Is this a problem with my router
> setup or with my ISP being too slow as it often lags and stalls
> doing normal Internet stuff like Web Surfing or FTP. If its a
> problem with my router could you give me some suggestions on
> how to fix it, the ICQ kernel patch wouldnt compile for me so
> thats not really an option unless you can tell me how to get it
> working under Slackware 4.0 (Kernel 2.2.6). Thankyou for your
> help, Alex.
If your using LICQ, your problems are there, it doesnt get
along well with chat. If not, you should probably raise your
ipchains MASQ timeouts. I use and reccomend: ipchains --
masquerade -M -S 360 360 360 It leaves the ports open, but it
keeps users from whining most of the time ;-) Best of luck!
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2430&cus_id=USENET&qtn_id=1588
------------------------------
From: "F.B.Quinn" <[EMAIL PROTECTED]>
Subject: Re: Linux Printing to a Remote Printer
Date: Tue, 10 Aug 1999 10:30:33 -0500
Reply-To: [EMAIL PROTECTED]
May I add an echo....?
I have the same problem "waiting for queue to be enabled on
<printername>.
I use Slakware for my client machines (486/33's) and Redhat 6 for my
servers (Pentium 100's), have read all the books, tried all the mantras
and burned incense. I am not using SAMBA, and have kept the installs
as austere as possible - ascii text only, opened up permissions,
and the maddening part is that some machines will and some machines
won't print remote. If I get it solved, I'll post the solution.
Thanks-
F. B. Quinn, Jr., M.D.
Univ.Tx Med. Branch
Galveston, TX
A
> The printcap file on the machine im trying to print from is:
>
> # Remote LasterJet IIp
> lp|lj|laser:\
> :sd=/var/spool/lpd/lj:\
> :rm=Printer:\
> :rp=lp7:\
> :lp=/dev/null:\
> :sh:
>
> I made the directory /var/spool/lpd/lj and currently it has chmod 666.
>
>
------------------------------
From: Mark Johnson <markj*no*spam*@gilanet.com>
Crossposted-To: comp.os.linux.setup,linux.sources.kernel
Subject: info about devpts filesystem desired
Date: Tue, 10 Aug 1999 13:05:59 -0600
Reply-To: markj*no*spam*@gilanet.com
Where can I find info on the devpts filesystem?
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: Q
Date: 10 Aug 1999 20:9:33 GMT
> Hi, I've got an interesting problem maybe you can help with:
> I have sucessfully set up an ftp server where the users are
> chroot'ed to their home directory. I have them set up with a
> bin, etc, lib, and incoming directory. The incoming directory
> is their depository for files uploading and downloading. It
> works great, but I want to put t wrinkle in this situation.
> You see, the ftp server is on a our Linux box and I want the
> users' "incoming" directory to reside on another box--our SGI
> fileserver. This did not present too much of a task and I can
> diagram how I did this. I created a shared directory on the
> SGI as /raid/ftpusers. Inside I have each user as a separate
> directory (ie. calvary, dark3, etc) with guest/user drwxrwxrwx
> access. I export the /raid/ftpusers via nfs and can mount it on
> the Linux box at a mountpoint /raid/ftpusers (named for
> consistency sake only). It has drwxrwxrwx access though it's
> root/root. As my user, I can cd to /raid/ftpusers/* and see
> anything and everything (my regular ftp users have a null
> shell). On the Linux side I have the users homes in
> /home2/users. Again they are chroot'ed to their home. Inside I
> have make a symbolic link to /raid/ftpusers/<user> as
> "incoming" (see below). Now when I 'ftp localhost' and log in
> as that user and try to cd into incoming, It says "550
> currwork: No such file or directory." Is it because of the
> chroot? The use of 'ln'? Access privilages? Just plain won't
> work that way? Here is the example: Internet || ---------
> -------------------- | |____
> /home2/users/<user> | LinuxBox |
> | ----------------------------- |-/bin |
> |-/lib |_ /raid/ftpusers (root/root drwxrwxrwx) |-/etc
> | |-/incoming <- ln -s
> NFS mount to: /raid/ftpusers/<user>
> | on Linux Box ---------------
> -------------- | | | SGI Server
> | | /raid/ftpusers | | | | |
> |- /<user> | | |- /<user> | |
> etc... | | (guest/user drwxrwxrwx) | -------------
> ---------------- Hope this clears things up of what I am
> trying to do. Any suggestions. I am wanting to do this
> because on the SGI, we have appletalk and wish to mount the
> SGI's /raid/ftpusers on the Mac as "ftpusers" (this is already
> done and works) so we can access and have backup of all ftp
> files without copying or moving or searching the Linuxbox for
> the latest files. This also keeps the Linuxbox from having
> space problems (the SGI has 70GB.) Anyhow, all suggestions
> welcome - except for insults ;-) If I have left anythign to
> question, ask me for clarity. Thanks -Wes Yates Belk Sent
> via Deja.com http://www.deja.com/ Share what you know. Learn
> what you don't.
Hi, chroot won't follow symbolic links, since it is located
under the root hierarchy (/). You will have to mount the
share within the same hierarchy ;( I would suggest you to get
a better ftp server. Pro Ftpd is good http://www.proftpd.org/,
that way you dont have to mess with libs and stuff that much.
(At least I saw patch once that did it) Hth!
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2469&cus_id=USENET&qtn_id=1579
------------------------------
From: QuestionExchange <[EMAIL PROTECTED]>
Subject: Re: IP-Number for device
Date: 10 Aug 1999 20:10:7 GMT
> Hi, I have to setup a dialin-server with isdn. Do I have to
> assign the same IP-Number on the ippp0- and eth0-device or do I
> need different IP-numbers? Thank You Martin Stockhammer
Hi, You can have both as the same. --vh
--
This answer is courtesy of QuestionExchange.com
http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2465&cus_id=USENET&qtn_id=1591
------------------------------
From: bizarrod <[EMAIL PROTECTED]>
Crossposted-To: redhat.networking.general
Subject: DHCP and RedHat 6.0, no gateway or DNS assigned???
Date: Tue, 10 Aug 1999 16:30:49 GMT
Anyone else having a problem in RedHat 6.0 using DHCP where you aren't
getting gateway and DNS assigned by DHCP??? I an using a redhat box for
ip masquerading for my cable modem. it worked fine with redhat 5.2, but
ever since i installed the new version, dhcp doesn't work right. I get an
ip address but no DNS addresses and no gateway address. I manually added
the DNS's and those never change so tht part isn't a big deal. However,
the gateway changes if I get a new ip address and then I have to spend a
few minutes tinkering with my startup files and reconfiguring the gateway.
Anyone else notice this problem??? Anyone know how to fix it?
Thanks
Bizarro Dave
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
** 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.networking) 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-Networking Digest
******************************
