November 2005 - openSUSE Users - openSUSE Mailing Lists

pppd doesn't always call ip-down
by Carlos E. R. 09 Dec '05

09 Dec '05

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, I have observed that, sometimes, when the modem connection goes down, the pppd daemon doesn't always call ip-down - or rather, not /etc/ppp/ip-down.local. The pppd daemon calls /etc/ppp/ip-up when the connection goes up, and /etc/ppp/ip-down when it goes down. Those scripts, written by SuSE, call in turn /etc/ppp/ip-up.local and /etc/ppp/ip-down.local, where we can place our customization. I employ those scripts (.local) to launch mail fetch/send cycles and a few things more, like clock sync. In fact, ip-down is a symlink to ip-up: there is only one script, but with two different sections based on the basename. The "down" section code of ip-down contains this: ip-down) restore_nameservers isdn_restartinterface start_firewall # call ip-down.local if it exists and is executable: if test -x /etc/ppp/ip-down.local ; then /etc/ppp/ip-down.local "$@" | logger -p security.notice -t ip-down.local > /dev/null & fi ... For example, tonight it failed. I can see that the pppd daemon calls ip-down: Nov 7 23:58:29 nimrodel pppd[7913]: Terminating connection due to lack of activity. Nov 7 23:58:29 nimrodel pppd[7913]: Connect time 4.4 minutes. Nov 7 23:58:29 nimrodel pppd[7913]: Sent 29179 bytes, received 144156 bytes. Nov 7 23:58:29 nimrodel pppd[7913]: Script /etc/ppp/ip-down started (pid 9097) Nov 7 23:58:29 nimrodel pppd[7913]: sent [LCP TermReq id=0x6 "Link inactive"] Nov 7 23:58:29 nimrodel pppd[7913]: rcvd [LCP TermAck id=0x6] Nov 7 23:58:29 nimrodel pppd[7913]: Connection terminated. Nov 7 23:58:29 nimrodel pppd[7913]: Waiting for 1 child processes... Nov 7 23:58:29 nimrodel pppd[7913]: script /etc/ppp/ip-down, pid 9097 * Nov 7 23:58:30 nimrodel pppd[7913]: Terminating on signal 15 * Nov 7 23:58:30 nimrodel pppd[7913]: sending SIGTERM to process 9097 * Nov 7 23:58:30 nimrodel pppd[7913]: Exit. Nov 7 23:58:29 nimrodel snort: pcap_loop: recvfrom: Network is down But ip-down does not run, it doesn't print anything. Normally, it would print something like this, at least: Nov 4 03:27:39 nimrodel ip-down: SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. _________________________^^^^^^^ but sometimes, like tonight, it doesn't print. Some times it runs, sometimes it doesn't. An intermittent failure... In fact, looking again at it, I think pppd is killing the script ip-down ("sending sigterm..."). Why? When it works, the log is somewhat different (the differences are marked with an *): Nov 7 21:52:41 nimrodel pppd[32132]: Terminating connection due to lack of activity. Nov 7 21:52:41 nimrodel pppd[32132]: Connect time 3.5 minutes. Nov 7 21:52:41 nimrodel pppd[32132]: Sent 42882 bytes, received 86405 bytes. Nov 7 21:52:41 nimrodel pppd[32132]: Script /etc/ppp/ip-down started (pid 649) Nov 7 21:52:41 nimrodel pppd[32132]: sent [LCP TermReq id=0x6 "Link inactive"] Nov 7 21:52:41 nimrodel pppd[32132]: rcvd [LCP TermAck id=0x6] Nov 7 21:52:41 nimrodel pppd[32132]: Connection terminated. Nov 7 21:52:41 nimrodel pppd[32132]: Waiting for 1 child processes... Nov 7 21:52:41 nimrodel pppd[32132]: script /etc/ppp/ip-down, pid 649 * Nov 7 21:52:44 nimrodel pppd[32132]: Script /etc/ppp/ip-down finished (pid 649), status = 0x0 Nov 7 21:52:44 nimrodel pppd[32132]: Exit. Where should I loot at? Ideas? (I have SuSE 9.3, but I have seen this behavior previously, I think). - -- Cheers, Carlos Robinson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFDb+wqtTMYHG2NR9URAgKIAKCTtA0+45JphEBe0igV5gCQpwbOBgCfajcd Qsq5AMZf/w4UQn1SFf4XNQs= =SAVJ -----END PGP SIGNATURE-----

1 2

PC running (SuSE 8.2 Pro / WinME) bootloader lilo -> grub
by Mick Higgins 08 Dec '05

08 Dec '05

Hi List, I have belatedly made the change from lilo to grub carefully converting my lilo config to grub'ese. It seems to work as expected. However, I have a couple of questions which I cannot track down in any off-line source (man, info, howto's etc.) and I am not able to trawl google or list archive with a time-limited unmetered dial-up connection at present. 1. In lilo I always used the boot-floppy option so as not to upset WinME. How do I do something similar in grub? 2. When the bootloader menu appeared in lilo I could interactively overide the default runlevel (ie: boot to runlevel 3 not 5) when necessary. How is this achieved in the grub menu? I will accept a URL to the appropriate list archive entry or web page. Apologies if this has been dealt with before. -- ** Message sent via KDE KMail v1.7.2 on SuSE Linux v8.2 ** Please reply in the format that this message is sent. ** Thanks.

6 5

Request for Codenames for SUSE Linux 10.1, 10.2
by Andreas Jaeger 07 Dec '05

07 Dec '05

During the 10.0 beta phase we had the theme "Towns of Development teams" and used these for Codenames like Prague, Boston etc. 10.1 Alpha1 was named "Auckland" under the same theme. For the future, I'd like to open the discussion on what we do. I propose the following policy: * We only use one Codename for the whole release, e.g. for the whole 10.1 release cycle, no codenames for snapshots or betas. * We have a theme that we can use for all our codenames, so the 10.1 and 10.2 codenames have the same theme. * A codename should be able to be visualized. I'd like to have e.g. a different photo on the grub boot screen for each Alpha/Beta etc. - all under the same codename. So, there should be different ways to visualize the object that are interesting and we can easily get say ten variations. What do you think of this? So, let's open a competition: Send your proposals for a theme together with some examples for codenames and how they can be visualized. An example: Theme: "Towns of Development Teams" Codenames: Prague, Boston, Nuernberg,... Visualization: Photo of a place in the town. I've created a wiki page for this, please add your proposals there, it is: http://www.opensuse.org/CodenameCompetition I'd like to choose 5 out of the proposals and then let's vote for the Codenames. Let's collect codenames for the next 7 days (until end of 2005-11-20) and then vote on them. Andreas -- Andreas Jaeger, aj(a)suse.de, http://www.suse.de/~aj SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126

12 31

isolinux initial keyboard dead (Asus P5GD2 Board, suse-10.0 CDs/DVDs)
by Bernd Melchers 06 Dec '05

06 Dec '05

Asus P5GD2-Premium Board Pentium 640 CPU, 2x512 MB Ram PS/2 Keyboard Cherry RS 6000 M ATi X600 Graphics Powercolor SATA Maxtor 250 GB disk I maintain some systems with the above configuration. They boot flawlessly with Knoppix, suse-9.2 and Windows XP-CDs. They work flawlessly with Knoppix, suse-9.2 and Windows XP. But it is NOT possible to boot them with (suse-9.3 or) Suse-10 installation or live CDs or DVDs, neither x64 nor i32 versions. I upgraded mb bios and DVD-Drive-Firmware, nothing helps. I changed BIOS PnP/OS, ACPI-2 and APIC Options, nothing helps. I see this: I insert the optical medium, boot, the Bonjour-Screen appears. Then when the grub boot menu appears, the keyboard is absolutely dead. So it is not possible to select an boot option. Therefore, after the time out, grub boots into the "boot from harddisk"-Menu-item and dies. I checked the MD5-Sum of the optical disks and could use them to install other computer with other mainboards. mit freundlichen Grüßen Bernd Melchers --

4 7

OpenOffice 2.0 and Gnome Menus
by Birgit Kellner 06 Dec '05

06 Dec '05

Hi, the OO 2.0 beta version delivered with SuSE 9.3 had proper Gnome file selector dialogues, but was otherwise buggy. I then installed the official OO 2.0 release from the source code as soon as it was available - gone were the nice Gtk dialogues, and I was left with the (much more impractical) original dialogues. Now I updated the whole system to OpenSuSe 10.0; again, no gnome dialogues, even though the gnome-plugins for OO are installed. Any clues how I can get the Gtk dialogues back? Thanks, Birgit Kellner

2 1

Memtest and installing 9.3
by JB 04 Dec '05

04 Dec '05

Hi gang, I'm trying to install 9.3 on my girlfriend's system but having some troubles. All the hardware seems to be picked up fine, and everything works great up until it's time to install all the software I picked. It starts to install, but at random intervals and (I guess) random packages, it pops up windows saying such-and-such rpm cannot be installed (something about bad checksum). We tried this 6 times (a LOT of time went by doing this). I later thought 'maybe it's the RAM'. I wrote her an e-mail telling her to do the memtest86 thing, but she says it opens a console and wants a root login name and root password. Since we haven't been able to install anything there is no root anything. How do I get the thing to do the memtest if there's no root name or password? I really want her to try Linux and she's wanting to also, even her daughter is pretty enthusiastic about trying it out. Thanks, John -- Ragheads and illegal aliens...the world's cockroaches.

5 7

Firewall stops internal ICMP.
by pelibali 04 Dec '05

04 Dec '05

Hi, I have a small home-net, where a SUSE 10.0 machine serves as a router. Everything works as expected, all the clients can surf / e-mail, whatever, but there is a smaller issue disturbing the harmony... We have dial-up connection and surprisingly I found, when we have no active connection, our clients _don't_ know about and they really wait until the initialized e.g. web-address will be timed-out. So in fact the router doesn't immediately let the clients know, that there is no connection and they have to find it out after a while, just "alone". Checking the firewall log showed me, that the ICMP (error-) messages don't arrive to the clients, because they get blocked (192.168.0.1 is the router, 192.168.0.6 is the client; in particular case trying to imap-ing e-mails from 146.123.123.123): Nov 26 11:28:17 trincsi kernel: SFW2-OUT-ERROR IN= OUT=eth0 SRC=192.168.0.1 DST=192.168.0.6 LEN=101 TOS=0x00 PREC=0xC0 TTL=64 ID=3105 PROTO=ICMP TYPE=3 CODE=0 [SRC=192.168.0.6 DST=146.123.123.123 LEN=73 TOS=0x00 PREC=0x00 TTL=64 ID=20782 DF PROTO=TCP SPT=59061 DPT=143 WINDOW=6368 RES=0x00 ACK PSH FIN URGP=0 OPT (0101080A0004BF9602DCDD03) ] Please could you suggest me a solution, how to get these ICMP packets arrive to my clients and let them through susefirewall? Maybe that's the same issue, but from clients I simply can't ping the router! I never needed it and don't plan to use something like that, but probably I would have to allow icmp somehow generally. I spent already hours with the configuration, until now no fruits. Any ideas are very welcome. Thank you, Pelibali

2 2

SuSE 10 Install Failure
by David Johanson 03 Dec '05

03 Dec '05

I purchased the boxed set of 10 and have been trying for over a month to get it installed on my SCSI based system. I put in a service request to Novell over a month ago and we have been "playing" around since then trying to see what's wrong with my system. When we finally reached a point of determining that the problem was that YaST was getting confused with the install because I have more than one CD/DVD drive in the system, it was announced, today, that Novell does not support installs on SCSI based systems anymore except on a pay for call basis and it was suggested I ask about the install problem on this list. So, has anyone had any problems installing on a SCSI based system and if so what has been the work around? System consists of the following: 4 matched removable SCSI HDs, only one of which is turned on during the install 1 each DVD writer and reader and 1 each CD writer and reader. All except the DVD writer are SCSI, the DVD writer is IDE and is the device used for the install attempts (I've tried all 4 devices but all fail during the loading of the packages) because it ( the IDE Sony DVD) is by far the fastest device. Below is a typical error I get when trying to perform the install during the package writing phase: "error: /var/adm/YaSTInstSrcManager/IS_CACHE_ 0X00000001/MEDIA/suse/i586/ghostscripts-fonts-other-8.15rc1-29.i586rpm:MD5 digest: BAD Expected (cb29d41081097a3cd4feeda767aeb453) != (e8aa68850260b19cbc0d44f43a330b2d)" This error will repeat with just a different package name for all packages other than the first three packages. Below is a quote from Novell tech service: "After installation of the first packages the access to /mnt/var/ gets lost, resulting in errors like: Could not open logfile '/mnt/var/log/YaST2/y2logRPM' cannot get exclusive lock on /mnt/var/lib/rpm/Packages cannot open Packages database in /mnt/var/lib/rpm Yast got confused by your quite unusual system." I've installed all versions of SuSE Linux, from 7.3 - 9.3 inclusive, on this box before without this problem. Any help on how to get 10 installed greatly appreciated. dave

10 23

SFW2-INext-DROP-DEFLT-INV in /var/log/warn?
by Steven T. Hatton 03 Dec '05

03 Dec '05

SFW2-INext-DROP-DEFLT-INV What does this mean? I'm having trouble transferring data from a site which I believe is very fast. I've noticed there are messages in the warn log containing the string shown above and also the IP address of the sever in question. Any ideas?

2 1

feature-request: Voting in Bugzilla
by Alexey Eremenko 03 Dec '05

03 Dec '05

Hi all! The current openSUSE project's bugzilla has 1 serious drawback - no voting system. Some problems might seems to be "small" by the development community, but turn out as serious issues for users. For this reason I ask you, the openSUSE project leaders to add a voting system to openSUSE bugzilla, like in Mandriva, where users can vote for problems/ bugs/feature-requests, that must be solved first.. A *Community* project needs this effective way to *communicate* with its users. -Alexey Eremenko. 20.nov.2005.

3 2