Menu
▾
▴
vacm-general
You can subscribe to this list here.
| 2000 |
Jan
(2) |
Feb
|
Mar
|
Apr
(3) |
May
|
Jun
|
Jul
(1) |
Aug
(26) |
Sep
(4) |
Oct
(7) |
Nov
(5) |
Dec
(19) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(7) |
Feb
(16) |
Mar
|
Apr
(12) |
May
(9) |
Jun
|
Jul
(23) |
Aug
(4) |
Sep
(8) |
Oct
(1) |
Nov
|
Dec
|
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2003 |
Jan
(2) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Matthew N. <mn...@us...> - 2001-02-01 16:44:09
|
San,
I originally tried the
vacm-2.0.0/nexxus/nexxus_modules/emp/support_utilities/ipmi_kcs_0.12.linux-2.2.14.patch
patch which the README file recommends for stock kernels (the one I'm
currently trying to patch is 2.2.18). The command "patch
-p1<ipmi_kcs_0.12.linux-2.2.14.patch" completed with a bunch of errors, so
yesterday I tried doing the same with the ipmi_kcs_1.0_linux-2.2.14.patch
and it patched fine. Unfortunately when I went to compile the kernel it
got a significant way through "make bzImage" before it died with:
cc -D__KERNEL__ -I/usr/src/linux-2.2.18-BG/include -E -C -P
-I/usr/src/linux-2.2
.18-BG/include -imacros
/usr/src/linux-2.2.18-BG/include/asm-i386/page_offset.h
-Ui386 arch/i386/vmlinux.lds.S >arch/i386/vmlinux.lds
ld -m elf_i386 -T /usr/src/linux-2.2.18-BG/arch/i386/vmlinux.lds -e stext
arch/i
386/kernel/head.o arch/i386/kernel/init_task.o init/main.o init/version.o \
--start-group \
arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o mm/mm.o
fs/f
s.o ipc/ipc.o \
fs/filesystems.a \
net/network.a \
drivers/block/block.a drivers/char/char.o drivers/misc/misc.a
drivers/ne
t/net.a drivers/scsi/scsi.a drivers/cdrom/cdrom.a drivers/pci/pci.a
drivers/pnp/
pnp.a drivers/video/video.a \
/usr/src/linux-2.2.18-BG/arch/i386/lib/lib.a
/usr/src/linux-2.2.18-BG/li
b/lib.a /usr/src/linux-2.2.18-BG/arch/i386/lib/lib.a \
--end-group \
-o vmlinux
drivers/char/char.o: In function `misc_init':
drivers/char/char.o(.text.init+0x438): undefined reference to
`ipmi_kcs_init'
make: *** [vmlinux] Error 1
#
At this point I'm not sure if this is something I'm doing wrong or if the
patch is causing problems. The system I'm running on is Red Hat 6.2.
Thanks for your help.
Matt Newton
IT Architect
IBM Global Services
T.J. Watson Research Center
San NeTTwerk Mehat <net...@va...> on 01/31/2001 01:41:37 AM
To: Matthew Newton/Watson/IBM@IBMUS
cc: vac...@li..., Mark E Giampapa/Watson/IBM@IBMUS,
Calin Cascaval/Watson/IBM@IBMUS
Subject: Re: [Vacm-general] using emp_extentd on linux kernels later than
2.2.14
what kinds of problems are you having?.. in a lot of cases there is a
little bit
of hand munging that may need to be done..
-san
On Wed, Jan 31, 2001 at 11:26:36AM -0500, Matthew Newton wrote:
> Hello,
>
> I've tried to apply the kernel patch for emp_extentd (in
>
vacm-2.0.0/nexxus/nexxus_modules/emp/support_utilities/ipmi_kcs_0.12.linux-2.2.14.patch)
>
> which the README file says should apply on stock non-VA Linux modified
> kernels, on both 2.2.15 and 2.2.18 kernels, without success. Has anyone
> gotten this to work with post-2.2.14 kernels, and if so, could they give
me
> some details on how they did it?
>
> Thanks,
>
> Matt Newton
> IBM Global Services
> T.J. Watson Research Center
>
>
> _______________________________________________
> Vacm-general mailing list
> Vac...@li...
> http://lists.sourceforge.net/lists/listinfo/vacm-general
--
-----------------------------------------------
San-Respect-my-.Xauthoritah-Mehat
Principal Engineer & Digital Groovemaster
VA Linux Systems
net...@va...
408-542-8654
-----------------------------------------------
|
|
From: San N. M. <net...@va...> - 2001-01-31 17:27:06
|
what kinds of problems are you having?.. in a lot of cases there is a little bit of hand munging that may need to be done.. -san On Wed, Jan 31, 2001 at 11:26:36AM -0500, Matthew Newton wrote: > Hello, > > I've tried to apply the kernel patch for emp_extentd (in > vacm-2.0.0/nexxus/nexxus_modules/emp/support_utilities/ipmi_kcs_0.12.linux-2.2.14.patch) > > which the README file says should apply on stock non-VA Linux modified > kernels, on both 2.2.15 and 2.2.18 kernels, without success. Has anyone > gotten this to work with post-2.2.14 kernels, and if so, could they give me > some details on how they did it? > > Thanks, > > Matt Newton > IBM Global Services > T.J. Watson Research Center > > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/lists/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Matthew N. <mn...@us...> - 2001-01-31 16:25:23
|
Hello, I've tried to apply the kernel patch for emp_extentd (in vacm-2.0.0/nexxus/nexxus_modules/emp/support_utilities/ipmi_kcs_0.12.linux-2.2.14.patch) which the README file says should apply on stock non-VA Linux modified kernels, on both 2.2.15 and 2.2.18 kernels, without success. Has anyone gotten this to work with post-2.2.14 kernels, and if so, could they give me some details on how they did it? Thanks, Matt Newton IBM Global Services T.J. Watson Research Center |
|
From: San N. M. <net...@va...> - 2001-01-26 15:30:33
|
Hey Matthew, Which version of VACM are you running? Are you using the old sercon_terminal program or just telnet? -san On Fri, Jan 26, 2001 at 10:28:22AM -0500, Matthew Newton wrote: > Hi Guys, > > I configured SERCON via COM1 on the 5 most critical machines in my cluster > without a problem. When, after a reboot, I tried to get access to the BIOS > Setup menu by hitting F2, using either telnet or sercon_terminal, it gets > interpreted as garbage (some set of 3-4 characters) and pretty much freaks > out the terminal from there. I'm wondering if there's a way to send it the > F2 keystroke that it will interpret correctly. > > Thanks for any help, > > Matt Newton > IBM Global Services > T.J. Watson Research Center > > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/lists/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Matthew N. <mn...@us...> - 2001-01-26 15:27:24
|
Hi Guys, I configured SERCON via COM1 on the 5 most critical machines in my cluster without a problem. When, after a reboot, I tried to get access to the BIOS Setup menu by hitting F2, using either telnet or sercon_terminal, it gets interpreted as garbage (some set of 3-4 characters) and pretty much freaks out the terminal from there. I'm wondering if there's a way to send it the F2 keystroke that it will interpret correctly. Thanks for any help, Matt Newton IBM Global Services T.J. Watson Research Center |
|
From: San N. M. <net...@va...> - 2001-01-02 15:35:38
|
Hey Matthew, happy new year... I have found that maintaining an EMp connection is an absolute lifesaver. Especially in co-located systems where you cannot readily access a system physically. The low level power control features as well as the 'whats going on? with the hardware?' features are indespensable... In a perfect world with EMP you hook up to BOTH serial ports... using bios , lilo, and linux console redirection over com1, while always having hardware control access on COM2 via EMP. Each topology is obviously different however, so its really up to you. Determine where the machines are going to be located.. are they remote? are they somewhere that you can get to them quickly should the unexpected happen? The machines for which you answer yes, i'd definetly make sure EMP was hooked up as *well* as serial console.... but at the very least, EMP... for the less critical machines, i'd reccommend just using the serial console stuff if that's important to you. If *all* machines are *equally* critical, then I reccommend getting more serial ports on your node controller. Either expand your digiboard out, add a rocketport, or whatnot.. depending on your application, downtime due to some unexpected failure of even a few hours can be as costly or often more costly then the price of a few extra serial ports... let me know if i can be of further service -san On Tue, Jan 02, 2001 at 09:48:26AM -0500, Matthew Newton wrote: > > Hi, > > The consensus seems to be that running a serial console tty through COM2 > while simultaneously using EMP (which is what I wanted to do) just isn't > possible. I was hoping this wasn't the case but I can certainly move > forward with setting up serial access to our linux cluster. Thanks very > much to all of you who responded so quickly - San, Dale, and Cameron - the > information you provided was extremely valuable and saved me a lot of time > and effort which I would have spent researching this myself. > > I guess that a follow-up question would be (and I'm sure this is subject to > individual opinion), if you have the choice between EMP and serial console > access for remote administration, which would you pick? My dilemma at the > moment is that I have enough serial ports from my Digi Board control > consoles to attach 1 serial connection to each node, plus 5-10 more (to use > for connections to COM1, if necessary). I'm leaning towards attaching EMP > to all of the nodes, and setting up console access via COM1 to the ~5 nodes > that are critical to the running of the system (e.g., the main NFS file > server, NIS masters & slaves, etc.). > > Happy New Year! > > Matt Newton > IBM Global Services > T.J. Watson Research Center > > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/mailman/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Matthew N. <mn...@us...> - 2001-01-02 14:47:35
|
Hi, The consensus seems to be that running a serial console tty through COM2 while simultaneously using EMP (which is what I wanted to do) just isn't possible. I was hoping this wasn't the case but I can certainly move forward with setting up serial access to our linux cluster. Thanks very much to all of you who responded so quickly - San, Dale, and Cameron - the information you provided was extremely valuable and saved me a lot of time and effort which I would have spent researching this myself. I guess that a follow-up question would be (and I'm sure this is subject to individual opinion), if you have the choice between EMP and serial console access for remote administration, which would you pick? My dilemma at the moment is that I have enough serial ports from my Digi Board control consoles to attach 1 serial connection to each node, plus 5-10 more (to use for connections to COM1, if necessary). I'm leaning towards attaching EMP to all of the nodes, and setting up console access via COM1 to the ~5 nodes that are critical to the running of the system (e.g., the main NFS file server, NIS masters & slaves, etc.). Happy New Year! Matt Newton IBM Global Services T.J. Watson Research Center |
|
From: San N. M. <net...@va...> - 2001-01-02 01:48:31
|
Hi, make sure you actualy *yhave* a com2 that is identified by the system. If this machine is equipped with an EMP (emergency management port) and you have it enabled, remember that com2 is used by the management processor and is NOT ACCESSABLE BY THE SYSTEM. If this is the case you have to do console stuff over COM1 and not COM2. On Thu, Dec 28, 2000 at 02:58:38PM -0500, Matthew Newton wrote: > Hi, > > I'm trying to get SERCON to work through the ttyS1/COM2 port on my > Intel-based managed nodes (running Red Hat 6.2). I have the following > entries in lilo.conf on the node to be managed: > > serial=1,19200n8 > append="console=ttyS1,19200" > > After commenting out the following line in /etc/inittab (just to avoid > confusion, don't know if this matters): > > #1:2345:respawn:/sbin/mingetty tty1 > > I added: > > s1:12345:respawn:/sbin/mingetty ttyS1 DT19200 vt100 > > I also did the following according to the VACM User's manual: > > cd /dev > rm -f console tty0 tty1 > mknod -m 622 console c 5 1 > mknod -m 622 tty0 c 4 0 > mknod -m 622 tty1 c 4 1 > > and, additionally, according to man ttys(4): > > mknod -m 660 /dev/ttyS1 c 4 65 > > Then deleted /etc/ioctl.save. > > Problem is when the machine boots I get the following in /var/log/messages: > > Dec 28 14:39:14 bg80 /sbin/mingetty[1211]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1212]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1213]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1214]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1215]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1216]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1217]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1218]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1219]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 /sbin/mingetty[1220]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 init: Id "s1" respawning too fast: disabled for 5 > minutes > > I think I've done everything right according to the VACM User Manual, but > for some reason Linux doesn't seem to think ttyS1 is configured right. > I've started reading through the Serial- and Text-Terminal HOWTOs, but was > hoping someone might recognize this before I have to slog through all the > manuals. Any help would be greatly appreciated. > > Thanks, > > Matt Newton > IBM Global Services > T.J. Watson Research Center > > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/mailman/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Dale H. <ro...@va...> - 2000-12-29 23:01:12
|
On Fri, Dec 29, 2000 at 08:33:45AM -0500, Matthew Newton wrote: > > The motherboard is an Intel L440GX+, and I was hoping that I could use the > EMP functionality in addition to the serial console functionality of the > SERCON module, using only the single COM2 port and a single cable instead > of the "recommended" cable to COM1 for SERCON and second cable to COM2 for > EMP. I didn't see anything in the VACM manual which said it wasn't > possible to use COM2 for both and a single cable. I have EMP turned on in > the BIOS, and additionally I have console redirect set in the BIOS to go > through COM2. If it wasn't possible to use both I would think these would > be mutually exclusive in the BIOS setup. I think it is pretty much mutually exclusive on the motherboard setup. True it isn't smart enough to figure out (in the Console Redirection area) that you may have dedicated the COM2 port to EMP. From the Intel doc's: 13. Select Com Port Address and select <2F8>, the address for COM2. EMP requires COM2 to function properly. ftp://download.intel.com/support/motherboards/server/l440gx/lgx_emp.pdf If you think about you are having a substantial amount of traffic going down one port. EMP shoots out data every 3 seconds or so. So if you are trying to use the same port for a serial console, you are going to get all kinds of garage on your screen. Not particular helpful. > > The entry: > > ttyS00 at 0x03f8 (irq = 4) is a 16550A > > appears in dmesg, so I guess tty0 is configured correctly whereas tty1 > isn't. Does turning on EMP somehow disable COM2 for use as a tty? > Yeah, makes it completely unavailable. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dale Harris <ro...@va...> VA Linux Systems Senior Support Engineer (temp. BSOP/SICE czar) 47071 Bayview Pkwy. (877) VA-LINUX Fremont, CA 94538 |
|
From: Matthew N. <mn...@us...> - 2000-12-29 22:37:47
|
Dale Harris wrote: >On Thu, Dec 28, 2000 at 02:58:38PM -0500, Matthew Newton wrote: >> Hi, >> >> I'm trying to get SERCON to work through the ttyS1/COM2 port on my >> Intel-based managed nodes (running Red Hat 6.2). I have the following >> entries in lilo.conf on the node to be managed: >> >> serial=1,19200n8 >> append="console=ttyS1,19200" > >I'm not entirely sure why you would want to do this. Are you not using the >EMP functionality of VACM? Or is this not a Intel motherboard that you >are running this on? The motherboard is an Intel L440GX+, and I was hoping that I could use the EMP functionality in addition to the serial console functionality of the SERCON module, using only the single COM2 port and a single cable instead of the "recommended" cable to COM1 for SERCON and second cable to COM2 for EMP. I didn't see anything in the VACM manual which said it wasn't possible to use COM2 for both and a single cable. I have EMP turned on in the BIOS, and additionally I have console redirect set in the BIOS to go through COM2. If it wasn't possible to use both I would think these would be mutually exclusive in the BIOS setup. > >> >> After commenting out the following line in /etc/inittab (just to avoid >> confusion, don't know if this matters): >> >> #1:2345:respawn:/sbin/mingetty tty1 >> >> I added: >> >> s1:12345:respawn:/sbin/mingetty ttyS1 DT19200 vt100 > >That should work. You might want to switch to /sbin/getty, just my personal >preference. > >Make sure you have an entry in /etc/securetty, so that you can actually log >into the machine on that device. I renamed the securetty file to something else to turn off the console login restriction on the nodes. I'm assuming that means I don't have to add tty1 in any access/configuration file. > >> Dec 28 14:39:14 bg80 /sbin/mingetty[1220]: /dev/ttyS1: not a tty >> Dec 28 14:39:14 bg80 init: Id "s1" respawning too fast: disabled for 5 >> minutes >> >> I think I've done everything right according to the VACM User Manual, but >> for some reason Linux doesn't seem to think ttyS1 is configured right. >> I've started reading through the Serial- and Text-Terminal HOWTOs, but was >> hoping someone might recognize this before I have to slog through all the >> manuals. Any help would be greatly appreciated. >> >> Thanks, >> >> Matt Newton >> IBM Global Services >> T.J. Watson Research Center > >What kind of motherboard are you trying to use this on? Is ttyS1 being >reported by the serial driver when the machine boot up (basically, do you >see it in `dmesg`)? If this is an Intel motherboard, and you are not >going to use EMP, then you will have to make sure that functionality is >turned off in the BIOS. The entry: ttyS00 at 0x03f8 (irq = 4) is a 16550A appears in dmesg, so I guess tty0 is configured correctly whereas tty1 isn't. Does turning on EMP somehow disable COM2 for use as a tty? > >-- > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >Dale Harris <ro...@va...> VA Linux Systems >Senior Support Engineer (temp. BSOP/SICE czar) 47071 Bayview Pkwy. >(877) VA-LINUX Fremont, CA 94538 > >_______________________________________________ >Vacm-general mailing list >Vac...@li... >http://lists.sourceforge.net/mailman/listinfo/vacm-general Thanks very much for your quick response, Dale. Matt Newton IBM Global Services T.J. Watson Research Center |
|
From: Cameron M. <cme...@va...> - 2000-12-29 22:37:32
|
On Thu, Dec 28, 2000 at 02:58:38PM -0500, Matthew Newton wrote: > Hi, > > I'm trying to get SERCON to work through the ttyS1/COM2 port on my > Intel-based managed nodes (running Red Hat 6.2). I have the following > entries in lilo.conf on the node to be managed: > > serial=1,19200n8 > append="console=ttyS1,19200" > > After commenting out the following line in /etc/inittab (just to avoid > confusion, don't know if this matters): > > #1:2345:respawn:/sbin/mingetty tty1 Leave this line. It is for the virtual console, not the real console. > I added: > > s1:12345:respawn:/sbin/mingetty ttyS1 DT19200 vt100 Minigetty in only for actual ttys. The ttyS1 is a serial port. Minigetty man pages say that it should not be used for serial ports. Use mgetty or getty. > > I also did the following according to the VACM User's manual: > > cd /dev > rm -f console tty0 tty1 > mknod -m 622 console c 5 1 > mknod -m 622 tty0 c 4 0 > mknod -m 622 tty1 c 4 1 > > and, additionally, according to man ttys(4): > > mknod -m 660 /dev/ttyS1 c 4 65 > > Then deleted /etc/ioctl.save. All this looks right. After you tell init to re-read the inittab check to see if a (m)getty process in running for ttyS1. > > I think I've done everything right according to the VACM User Manual, but > for some reason Linux doesn't seem to think ttyS1 is configured right. > I've started reading through the Serial- and Text-Terminal HOWTOs, but was > hoping someone might recognize this before I have to slog through all the > manuals. Any help would be greatly appreciated. I noticed that the instructions said to use minigetty and ment to say something. Whoever is in charge of the Docs, cuold you look at this? I think mingetty should be mgetty or just getty. cameron -- Cameron Meadors cme...@va... Professional Services Engineer Tel. 408.543.8760 VA Linux Cell. 408.887.3181 |
|
From: Dale H. <ro...@va...> - 2000-12-29 03:59:00
|
On Thu, Dec 28, 2000 at 02:58:38PM -0500, Matthew Newton wrote: > Hi, > > I'm trying to get SERCON to work through the ttyS1/COM2 port on my > Intel-based managed nodes (running Red Hat 6.2). I have the following > entries in lilo.conf on the node to be managed: > > serial=1,19200n8 > append="console=ttyS1,19200" I'm not entirely sure why you would want to do this. Are you not using the EMP functionality of VACM? Or is this not a Intel motherboard that you are running this on? > > After commenting out the following line in /etc/inittab (just to avoid > confusion, don't know if this matters): > > #1:2345:respawn:/sbin/mingetty tty1 > > I added: > > s1:12345:respawn:/sbin/mingetty ttyS1 DT19200 vt100 That should work. You might want to switch to /sbin/getty, just my personal preference. Make sure you have an entry in /etc/securetty, so that you can actually log into the machine on that device. > Dec 28 14:39:14 bg80 /sbin/mingetty[1220]: /dev/ttyS1: not a tty > Dec 28 14:39:14 bg80 init: Id "s1" respawning too fast: disabled for 5 > minutes > > I think I've done everything right according to the VACM User Manual, but > for some reason Linux doesn't seem to think ttyS1 is configured right. > I've started reading through the Serial- and Text-Terminal HOWTOs, but was > hoping someone might recognize this before I have to slog through all the > manuals. Any help would be greatly appreciated. > > Thanks, > > Matt Newton > IBM Global Services > T.J. Watson Research Center What kind of motherboard are you trying to use this on? Is ttyS1 being reported by the serial driver when the machine boot up (basically, do you see it in `dmesg`)? If this is an Intel motherboard, and you are not going to use EMP, then you will have to make sure that functionality is turned off in the BIOS. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dale Harris <ro...@va...> VA Linux Systems Senior Support Engineer (temp. BSOP/SICE czar) 47071 Bayview Pkwy. (877) VA-LINUX Fremont, CA 94538 |
|
From: Matthew N. <mn...@us...> - 2000-12-29 02:59:16
|
Hi, I'm trying to get SERCON to work through the ttyS1/COM2 port on my Intel-based managed nodes (running Red Hat 6.2). I have the following entries in lilo.conf on the node to be managed: serial=1,19200n8 append="console=ttyS1,19200" After commenting out the following line in /etc/inittab (just to avoid confusion, don't know if this matters): #1:2345:respawn:/sbin/mingetty tty1 I added: s1:12345:respawn:/sbin/mingetty ttyS1 DT19200 vt100 I also did the following according to the VACM User's manual: cd /dev rm -f console tty0 tty1 mknod -m 622 console c 5 1 mknod -m 622 tty0 c 4 0 mknod -m 622 tty1 c 4 1 and, additionally, according to man ttys(4): mknod -m 660 /dev/ttyS1 c 4 65 Then deleted /etc/ioctl.save. Problem is when the machine boots I get the following in /var/log/messages: Dec 28 14:39:14 bg80 /sbin/mingetty[1211]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1212]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1213]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1214]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1215]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1216]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1217]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1218]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1219]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 /sbin/mingetty[1220]: /dev/ttyS1: not a tty Dec 28 14:39:14 bg80 init: Id "s1" respawning too fast: disabled for 5 minutes I think I've done everything right according to the VACM User Manual, but for some reason Linux doesn't seem to think ttyS1 is configured right. I've started reading through the Serial- and Text-Terminal HOWTOs, but was hoping someone might recognize this before I have to slog through all the manuals. Any help would be greatly appreciated. Thanks, Matt Newton IBM Global Services T.J. Watson Research Center |
|
From: Zac S. <zsp...@va...> - 2000-12-21 19:20:04
|
On Thu, Dec 21, 2000 at 12:08:36PM -0700, Nelson, Ben wrote: > I'm having trouble connecting to my nexxus remotely. I keep getting the > following error: > > vash$ connect hostname > vash: Unable to ping nexxus at hostname > > I wasn't sure if this 'ping' was referring to a standard ICMP ping or if the > word ping means something else in vacm. It's not a firewalling issue, > because for testing purposes, I opened the firewall up, allowing all traffic > from my machine to my vacm controller. I can connect to the nexxus no > problem when I am logged into the controller. But, ultimately, I'd like to > be able to run flim from my machine and have it connect to the nexxus > remotely. That means that vash can't send the ping command to the nexxus at host "hostname". Make sure the port is not firewalled and that the machine is indeed running a nexxus. Zac -- VACM - Couldn't Your SysAdmin Use Some Beauty Sleep? http://vacm.sourceforge.net |
|
From: Nelson, B. <bn...@ri...> - 2000-12-21 19:12:25
|
I'm having trouble connecting to my nexxus remotely. I keep getting the following error: vash$ connect hostname vash: Unable to ping nexxus at hostname I wasn't sure if this 'ping' was referring to a standard ICMP ping or if the word ping means something else in vacm. It's not a firewalling issue, because for testing purposes, I opened the firewall up, allowing all traffic from my machine to my vacm controller. I can connect to the nexxus no problem when I am logged into the controller. But, ultimately, I'd like to be able to run flim from my machine and have it connect to the nexxus remotely. --Ben ---------------------------------------------- Ben Nelson Hosting Systems Administrator RightNow Technologies bn...@ri... |
|
From: Dale H. <ro...@va...> - 2000-12-20 19:03:59
|
On Wed, Dec 20, 2000 at 08:04:38AM -0800, Ian S. McLeod wrote: > -Kernel boot messages redirect _incorrectly_ (appears to be at 9600 baud) Well, you probably checked this, but is it actually set to 19200 in the BIOS? I wonder also maybe some serial configuration somewhere else might be getting in the way. I know on Slack you have a /etc/serial.conf, but I don't think RedHat distros use that. Dale |
|
From: San N. M. <net...@va...> - 2000-12-20 17:08:35
|
hey ian, as far as baudrate is concerned, if you get ANY data at the correct baudrate then sercon is working properly. the problem MUST be on the remote side. sercon NEVER changes its baud rate.. whatever you set it to is what it uses. so if you're getting BIOS and LILO correctly, but after that the bps looks wrong, then its system side and not sercon as sercon has not done anything to change the baudrate inbetween. ill look at the password echo thing next week -san On Wed, Dec 20, 2000 at 08:04:38AM -0800, Ian S. McLeod wrote: > Hello all, > > I'm seeing two strange problems when using sercon_terminal with some VA > Linux 2200 systems. > > 1) Baud rate problems > > I have _all_ relevant files set to use 19200 baud. To be specific: > > -lilo.conf has serial redirect set > -lilo.conf has console= at the correct baud rate for the default boot > image > -the bios redirect is set > -inittab has a serial console set at the correct baud rate > > I have also deleted /etc/ioctl.save _many_ times. I consistently observe > the following: > > -BIOS redirects correctly > -LILO prompt redirects correctly > -Kernel boot messages redirect _incorrectly_ (appears to be at 9600 baud) > -Login prompt redirects correctly > > > 2) Character echo > > On these same systems we reliably see the password echo back when logging > in. > > Any ideas? > > -Ian > > > > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/mailman/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Ian S. M. <ia...@va...> - 2000-12-20 16:04:41
|
Hello all, I'm seeing two strange problems when using sercon_terminal with some VA Linux 2200 systems. 1) Baud rate problems I have _all_ relevant files set to use 19200 baud. To be specific: -lilo.conf has serial redirect set -lilo.conf has console= at the correct baud rate for the default boot image -the bios redirect is set -inittab has a serial console set at the correct baud rate I have also deleted /etc/ioctl.save _many_ times. I consistently observe the following: -BIOS redirects correctly -LILO prompt redirects correctly -Kernel boot messages redirect _incorrectly_ (appears to be at 9600 baud) -Login prompt redirects correctly 2) Character echo On these same systems we reliably see the password echo back when logging in. Any ideas? -Ian |
|
From: Scott S. <ss...@va...> - 2000-12-14 18:24:42
|
Anyone know what this means? IPMI KCS driver (San Mehat net...@va...) v1.2 at io 0xca2 --Using 4096 byte ring buffer for incomming IPMI data --Using 4096 byte ring buffer for board FRU --Polling frequency set to 10 ticks [IPMI_KCS] WARNING: Out of Band data flag set (data 0x8a) [IPMI_KCS] Email net...@va... --BMC watchdog explicitly disabled --FRU download commencing [IPMI_KCS] WARNING: Out of Band data flag set (data 0x8b) [IPMI_KCS] Email net...@va... [IPMI_KCS] Language 0xb6 dependant decode not supported [IPMI_KCS] Language 0xb6 dependant decode not supported [IPMI_KCS] Six bit ASCII decode not supported [IPMI_KCS] Language 0xb6 dependant decode not supported --FRU download complete [IPMI_KCS] WARNING: Out of Band data flag set (data 0x8b) [IPMI_KCS] Email net...@va... [IPMI_KCS] BAD ISA state [IPMI_KCS] Read Message failed Thanks in advance for any help. -Scott -- Scott Sawyer Professional Services Engineer VALinux Systems - http://www.valinux.com ph: (602) 920-0083 |
|
From: San N. M. <net...@va...> - 2000-12-12 01:42:06
|
you'll also get it when you havent configured the emp module On Mon, Dec 11, 2000 at 05:07:06PM -0800, Zac Sprackett wrote: > On Mon, Dec 11, 2000 at 03:00:04AM -0800, San NeTTwerk Mehat wrote: > > You'll get that message if you havent configured the EMP module with a serial > > port, etc for EMP traffic. basically the EMP module is disabled > > Actually, I beleive the issue is modmapping. Modmmapping was added just after > we released 2.0.0. This allows you to specify which modules to load for a > given node. > > Try sending the following ipc from vash: > > ipc localhost nexxus:node_modmap_add:<NODE NAME>:EMP > > You will need to enable each module you plan on using. > > Zac > -- > VACM - Couldn't Your SysAdmin Use Some Beauty Sleep? > > http://vacm.sourceforge.net > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/mailman/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Dean J. <dt...@fu...> - 2000-12-12 01:34:00
|
Zac Sprackett wrote: > > On Mon, Dec 11, 2000 at 03:00:04AM -0800, San NeTTwerk Mehat wrote: > > You'll get that message if you havent configured the EMP module with a serial > > port, etc for EMP traffic. basically the EMP module is disabled > > Actually, I beleive the issue is modmapping. Modmmapping was added just after > we released 2.0.0. This allows you to specify which modules to load for a > given node. > > Try sending the following ipc from vash: > > ipc localhost nexxus:node_modmap_add:<NODE NAME>:EMP > > You will need to enable each module you plan on using. > I thought that the modmapping stuff defaults to everything if nothing is specified. That was something that I had run into. -Dean |
|
From: Zac S. <zsp...@va...> - 2000-12-12 01:07:07
|
On Mon, Dec 11, 2000 at 03:00:04AM -0800, San NeTTwerk Mehat wrote: > You'll get that message if you havent configured the EMP module with a serial > port, etc for EMP traffic. basically the EMP module is disabled Actually, I beleive the issue is modmapping. Modmmapping was added just after we released 2.0.0. This allows you to specify which modules to load for a given node. Try sending the following ipc from vash: ipc localhost nexxus:node_modmap_add:<NODE NAME>:EMP You will need to enable each module you plan on using. Zac -- VACM - Couldn't Your SysAdmin Use Some Beauty Sleep? http://vacm.sourceforge.net |
|
From: San N. M. <net...@va...> - 2000-12-11 21:45:10
|
You'll get that message if you havent configured the EMP module with a serial port, etc for EMP traffic. basically the EMP module is disabled -san On Mon, Dec 11, 2000 at 10:28:48AM -0800, Mike Sklar wrote: > > Hi, > > I recently decided to rebuild my master console node with redhat 7.0 so that > it could run my companies internal build. I originally was running beta 2.0r4 > and today checked out 2.1.0devel1 from the sourceforge CVS tree. Built great > and I installed on my master node. Unfortunatly I am encountering the same > problem with both versions. Whenever I go to use a emp modules function I > get "EMP:6:JOB_ERROR:Package not installed" message. > > vash$ ipc localhost emp:node_status:bigfoot4 > EMP:6:JOB_STARTED > EMP:6:JOB_ERROR:Package not installed > > Now I am able to use sysstat, sercon, and nexxus without any issues, but > there seems to be some sort of system misconfiguration on my part that > I am running into the emp module message. Quering the nexxus module says that > emp is present and working. Any ideas? > > vash$ ipc localhost nexxus:nexxus_modules > NEXXUS:11:JOB_STARTED > NEXXUS:11:MODULE:VA1000:VA 1000 Module:Jerry Katzung > (ka...@va...):Module to support VA Linux 1000 > clusters.:1.1:VA1000:STATE_RUNNING > NEXXUS:11:MODULE:BayTech:Baytech Power Strip Management Module:Zac > Sprackett (zsp...@va...):Provides VACM interface to the Baytech > RPC series of managed power strips:2.0:BAYTECH:STATE_RUNNING > NEXXUS:11:MODULE:Sysstat:System Status Daemon Module:San Mehat > (net...@va...):Module to interface with a remote VACM System > Status Daemon:0.1:SYSSTAT:STATE_RUNNING > NEXXUS:11:MODULE:APCUPS:APC SmartUPS(tm) module:San Mehat > (net...@va...):Module to communicate with APC > SmartUPS(tm) hardware:1.0:APCUPS:STATE_RUNNING > NEXXUS:11:MODULE:EMP:Emergency Management Port Module:San Mehat > (net...@va...):Module to interface with a remote Emergency > Management Port:2.0:EMP:STATE_RUNNING > NEXXUS:11:MODULE:msc:Remote Services Module:Dean Johnson > (dt...@sg...):Module to perform remote tasks using > MSC:2.0:MSC:STATE_RUNNING > NEXXUS:11:MODULE:SERCON:Remote Serial Console Module:San Mehat > (net...@va...):Module to interface with a remote serial > console:2.0:SERCON:STATE_RUNNING > NEXXUS:11:MODULE:ICMP ECHO:ICMP Echo Tester:San Mehat > (net...@va...):Sends ICMP ECHO messages to a remote host to test > connectivity:1.0:ICMP_ECHO:STATE_RUNNING > NEXXUS:11:MODULE:rsh:Remote Services Module:Dean Johnson (dt...@sg...) & > Zac Sprackett (zsp...@va...):Module to perform remote tasks > using RSH or SSH:2.0:RSH:STATE_RUNNING > NEXXUS:11:JOB_COMPLETED > > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/mailman/listinfo/vacm-general -- ----------------------------------------------- San-Respect-my-.Xauthoritah-Mehat Principal Engineer & Digital Groovemaster VA Linux Systems net...@va... 408-542-8654 ----------------------------------------------- |
|
From: Mike S. <sk...@sc...> - 2000-12-11 19:03:39
|
> Do you run VA1000 systems or the older 2U systems that use the rocketport card? 2u the 2200 systems. > If you use the rocketport card, /etc/sysconfig/vacm must contain the following: > ROCKET=true > Setting this variable true tells the script /etc/rc.d/init.d/vacm to load the > rocketport module into the kernel before starting nexxus. asdfjkl$ rsh bigfoot0 cat /etc/sysconfig/vacm ROCKET=true Yes, I have had ROCKET=true in place all along. |
|
From: Thomas A. Y. <ta...@bn...> - 2000-12-11 18:46:38
|
Do you run VA1000 systems or the older 2U systems that use the rocketport card? If you use the rocketport card, /etc/sysconfig/vacm must contain the following: ROCKET=true Setting this variable true tells the script /etc/rc.d/init.d/vacm to load the rocketport module into the kernel before starting nexxus. I hope this helps. ---------------------- Thomas Yanuklis ta...@bn... Mike Sklar wrote: > Hi, > > I recently decided to rebuild my master console node with redhat 7.0 so that > it could run my companies internal build. I originally was running beta 2.0r4 > and today checked out 2.1.0devel1 from the sourceforge CVS tree. Built great > and I installed on my master node. Unfortunatly I am encountering the same > problem with both versions. Whenever I go to use a emp modules function I > get "EMP:6:JOB_ERROR:Package not installed" message. > > vash$ ipc localhost emp:node_status:bigfoot4 > EMP:6:JOB_STARTED > EMP:6:JOB_ERROR:Package not installed > > Now I am able to use sysstat, sercon, and nexxus without any issues, but > there seems to be some sort of system misconfiguration on my part that > I am running into the emp module message. Quering the nexxus module says that > emp is present and working. Any ideas? > > vash$ ipc localhost nexxus:nexxus_modules > NEXXUS:11:JOB_STARTED > NEXXUS:11:MODULE:VA1000:VA 1000 Module:Jerry Katzung > (ka...@va...):Module to support VA Linux 1000 > clusters.:1.1:VA1000:STATE_RUNNING > NEXXUS:11:MODULE:BayTech:Baytech Power Strip Management Module:Zac > Sprackett (zsp...@va...):Provides VACM interface to the Baytech > RPC series of managed power strips:2.0:BAYTECH:STATE_RUNNING > NEXXUS:11:MODULE:Sysstat:System Status Daemon Module:San Mehat > (net...@va...):Module to interface with a remote VACM System > Status Daemon:0.1:SYSSTAT:STATE_RUNNING > NEXXUS:11:MODULE:APCUPS:APC SmartUPS(tm) module:San Mehat > (net...@va...):Module to communicate with APC > SmartUPS(tm) hardware:1.0:APCUPS:STATE_RUNNING > NEXXUS:11:MODULE:EMP:Emergency Management Port Module:San Mehat > (net...@va...):Module to interface with a remote Emergency > Management Port:2.0:EMP:STATE_RUNNING > NEXXUS:11:MODULE:msc:Remote Services Module:Dean Johnson > (dt...@sg...):Module to perform remote tasks using > MSC:2.0:MSC:STATE_RUNNING > NEXXUS:11:MODULE:SERCON:Remote Serial Console Module:San Mehat > (net...@va...):Module to interface with a remote serial > console:2.0:SERCON:STATE_RUNNING > NEXXUS:11:MODULE:ICMP ECHO:ICMP Echo Tester:San Mehat > (net...@va...):Sends ICMP ECHO messages to a remote host to test > connectivity:1.0:ICMP_ECHO:STATE_RUNNING > NEXXUS:11:MODULE:rsh:Remote Services Module:Dean Johnson (dt...@sg...) & > Zac Sprackett (zsp...@va...):Module to perform remote tasks > using RSH or SSH:2.0:RSH:STATE_RUNNING > NEXXUS:11:JOB_COMPLETED > > _______________________________________________ > Vacm-general mailing list > Vac...@li... > http://lists.sourceforge.net/mailman/listinfo/vacm-general |