VMware Crashing with Suse8
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Install and config of VMware 3.2 is proving to be a nightmare. I found this on Google:
VMware has a problem with Suse 8.0.
Suse decided to include a memory management patch from the 2.5.x kernel tree in their 2.4.x kernels. Applications that are not specifically modified for this patch can be severely broken by it.
VMware was not written to take this patch into account because VMware only supports 2.2.x and 2.4.x kernels on the host and this patch is *NOT* part of the 2.2.x or 2.4.x kernels.
Suse should not be claiming that the kernel shipped with Suse 8.0 is a 2.4.18 kernel. It is not. It is a heavily modified kernel that was originally based on 2.4.18. VMware is compatible with 2.4.18 kernels. It is not 100% compatible with the custom kernel installed by default in Suse 8.0 (please note that Suse 8.0 is not a supported host OS)
Replacing your patched Suse kernel with a stock 2.4.18 kernel will resolve the problem completely.
a fast an simple solution is to limit RAM to 850 MB at bootime, works but if you need the RAM....
Whether you need the RAM or not, the better solution is to remove that patch. Apparently even the author of the patch has said that the only people who should see any benefit from it right now are people with more than 1GB of RAM, but less than 2GB of RAM who are also running Oracle on the machine. So removing it should not hurt anything.
So is the only fix to use kernel.org kernels? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2XoIIACgkQnQ18+PFcZJtREQCfXcbD8Bfueuky9/0dAe6vH8NX 55cAniZA0uVO5sDKf0La3twX02wKv+Q+ =aqb7 -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 This is a deal-killer for me. When I install and config VMware the whole system goes unstable, particularly networking. For some reason VMware insists on putting both HostOnly and NAT on vmnet8, which thereafter hangs the network and completely precludes a clean system shutdown. vmware-config.pl cannot shut down networking services once started and Aborts, so I get one shot per hang/powerdown/boot. As the system is unstable this way, there's no tinkering with VMware. So I config VMware with NO networking and I can then config and reboot, but when I power on the VM... Segfault (no Oops). I have an IBM A22p notebook with 512MB of RAM, and Suse8/K3.0.3/kernel2.4.18-154, with all updates. Google wisdom seems to be the problem is a patch in Suse's kernel which enables 4GB of RAM. Suse's workaround ( http://sdb.suse.de/en/sdb/html/fhassel_vmware_segfault.html ) is to add to lilo.conf append = "mem=512M" and lilo. But it doesn't help for me. Googlers say to ditch the Suse kernel, or at least turn off upper memory and recompile. Well the kernel compile went fine, but make modules crashes with some arcane error that I just don't have time to decipher. VMware 3.2 says one of its new features is 'support for Suse8', which is interesting. But I've tried every vmware tar and rpm available (including stock one from Suse8) and all have the same problems. So, will Suse assist here? This is mission-critical for us. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2YXaMACgkQnQ18+PFcZJueLQCfT7oPmzOYXV0JW4MafhWZbrQ4 QAEAn13OdyOK5WSQBY+CqNBCf6u+N9Yk =0QQC -----END PGP SIGNATURE-----
Carl said:
So, will Suse assist here? This is mission-critical for us.
Double-wrong forum. Firstly it has little or nothing to do with kde, so suse-kde is definitely wrong, and the mailing lists in general isn't suse's official support forum. If a suser reads it it's more luck than anything. One of the official support addresses you can find on SuSE's homepage would be the best idea Incidentally, I run vmware 3.1.1 with no problems whatever. Perhaps that's an acceptable alternative for you //Anders
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Monday, 30 September 2002 9:32, Anders Johansson wrote:
Carl said:
So, will Suse assist here? This is mission-critical for us.
Double-wrong forum. Firstly it has little or nothing to do with kde, so suse-kde is definitely wrong, and the mailing lists in general isn't suse's official support forum. If a suser reads it it's more luck than anything. One of the official support addresses you can find on SuSE's homepage would be the best idea
That's aall very nice, but Suse does not answer emails. And the main list gives 200 messages a day, which most mortals cannot sift. I don't want to add the load to my ISP, just to filter out 95% of the mail.
Incidentally, I run vmware 3.1.1 with no problems whatever. Perhaps that's an acceptable alternative for you
I've not tried 3.1.1, but will. Although they say on Google that it has the same problems. Did you try 3.2 and fail? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2YZQYACgkQnQ18+PFcZJul3wCfWfyZx/m/qUHgRR2uv2f9TL43 BqUAniPk+DjVB87XXE4bn1oN85zi7o6A =5Mq2 -----END PGP SIGNATURE-----
Carl said:
I've not tried 3.1.1, but will. Although they say on Google that it has the same problems. Did you try 3.2 and fail?
No, I ran 3.1.1 from the start and since I haven't seen a need to upgrade I haven't bothered to check if any are available
Carl, None of the problems you describe look like SuSE; they seem to be VMware related. Shouldn't you try their tech support and lists? If your business is affected that seriously you should consider SuSE's "Paid Support" services. Best of Luck ............ PeterB On Monday 30 September 2002 09:20, Carl wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
This is a deal-killer for me.
When I install and config VMware the whole system goes unstable, particularly networking. For some reason VMware insists on putting both HostOnly and NAT on vmnet8, which thereafter hangs the network and completely precludes a clean system shutdown. vmware-config.pl cannot shut down networking services once started and Aborts, so I get one shot per hang/powerdown/boot. As the system is unstable this way, there's no tinkering with VMware.
So I config VMware with NO networking and I can then config and reboot, but when I power on the VM... Segfault (no Oops).
I have an IBM A22p notebook with 512MB of RAM, and Suse8/K3.0.3/kernel2.4.18-154, with all updates.
Google wisdom seems to be the problem is a patch in Suse's kernel which enables 4GB of RAM. Suse's workaround ( http://sdb.suse.de/en/sdb/html/fhassel_vmware_segfault.html ) is to add to lilo.conf append = "mem=512M" and lilo. But it doesn't help for me.
Googlers say to ditch the Suse kernel, or at least turn off upper memory and recompile. Well the kernel compile went fine, but make modules crashes with some arcane error that I just don't have time to decipher.
VMware 3.2 says one of its new features is 'support for Suse8', which is interesting. But I've tried every vmware tar and rpm available (including stock one from Suse8) and all have the same problems.
So, will Suse assist here? This is mission-critical for us. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org
iEYEARECAAYFAj2YXaMACgkQnQ18+PFcZJueLQCfT7oPmzOYXV0JW4MafhWZbrQ4 QAEAn13OdyOK5WSQBY+CqNBCf6u+N9Yk =0QQC -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Well I've been compiling my red arse off, and think I have a workaround. VMware is behind, and their new release doesn't fix this, as it purports. But now when I boot I get the error message: <notice>/etc/init.d/rc5.d/S13alsasound start Starting sound driver<notice>'/etc/init.d/rc5.d/S13alsasound start' exits with status 0 : snd_cs46xxmodprobe: Can't locate module snd_cs46xx failed A check in modules.conf shows indeed that snd_cs46xx is being called, but the module is actually cs46xx. So I presume there is some invisible massaging, that has now broken. I did disable most sound modules, leaving only the generic ones, Soundblaster, Crystal, and OPL. Anyone have a clue? On Monday, 30 September 2002 11:15, you wrote:
Carl, None of the problems you describe look like SuSE; they seem to be VMware related. Shouldn't you try their tech support and lists?
If your business is affected that seriously you should consider SuSE's "Paid Support" services.
Best of Luck ............ PeterB
On Monday 30 September 2002 09:20, Carl wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
This is a deal-killer for me.
When I install and config VMware the whole system goes unstable, particularly networking. For some reason VMware insists on putting both HostOnly and NAT on vmnet8, which thereafter hangs the network and completely precludes a clean system shutdown. vmware-config.pl cannot shut down networking services once started and Aborts, so I get one shot per hang/powerdown/boot. As the system is unstable this way, there's no tinkering with VMware.
So I config VMware with NO networking and I can then config and reboot, but when I power on the VM... Segfault (no Oops).
I have an IBM A22p notebook with 512MB of RAM, and Suse8/K3.0.3/kernel2.4.18-154, with all updates.
Google wisdom seems to be the problem is a patch in Suse's kernel which enables 4GB of RAM. Suse's workaround ( http://sdb.suse.de/en/sdb/html/fhassel_vmware_segfault.html ) is to add to lilo.conf append = "mem=512M" and lilo. But it doesn't help for me.
Googlers say to ditch the Suse kernel, or at least turn off upper memory and recompile. Well the kernel compile went fine, but make modules crashes with some arcane error that I just don't have time to decipher.
VMware 3.2 says one of its new features is 'support for Suse8', which is interesting. But I've tried every vmware tar and rpm available (including stock one from Suse8) and all have the same problems.
So, will Suse assist here? This is mission-critical for us. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org
iEYEARECAAYFAj2YXaMACgkQnQ18+PFcZJueLQCfT7oPmzOYXV0JW4MafhWZbrQ4 QAEAn13OdyOK5WSQBY+CqNBCf6u+N9Yk =0QQC -----END PGP SIGNATURE-----
- -- Remember, drive defensively! And of course, the best defense is a good offense! -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2YnbUACgkQnQ18+PFcZJtu6wCdEobTkK3jYIEBVsV1HejFeLnr ZnoAn3io3LyPDAJ/zOKNAFERgQ+o6p/x =3HoL -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I guess Suse is not going to step in here. This problem has taken at least 12 months off my life. In addition to the sound driver problem on boot with my custom kernel, when trying to compile modules, it fails on the TI 1394 driver: make[2]: Entering directory `/usr/src/linux-2.4.18.SuSE/drivers/ieee1394' gcc -D__KERNEL__ -I/usr/src/linux-2.4.18.SuSE/include -Wall - -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer - -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 - -march=i686 -DMODULE -DKBUILD_BASENAME=pcilynx -c -o pcilynx.o pcilynx.c pcilynx.c: In function `mem_open': pcilynx.c:647: `num_of_cards' undeclared (first use in this function) pcilynx.c:647: (Each undeclared identifier is reported only once pcilynx.c:647: for each function it appears in.) pcilynx.c:647: `cards' undeclared (first use in this function) pcilynx.c: In function `aux_poll': pcilynx.c:706: `cards' undeclared (first use in this function) make[2]: *** [pcilynx.o] Error 1 make[2]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers/ieee1394' make[1]: *** [_modsubdir_ieee1394] Error 2 make[1]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers' make: *** [_mod_drivers] Error 2 hydra:/usr/src/linux # This is something we need. As I am not a C programmer, I cannot know what is wrong. Does anyone see what's wrong? On Monday, 30 September 2002 13:53, you wrote:
Well I've been compiling my red arse off, and think I have a workaround. VMware is behind, and their new release doesn't fix this, as it purports.
But now when I boot I get the error message:
<notice>/etc/init.d/rc5.d/S13alsasound start Starting sound driver<notice>'/etc/init.d/rc5.d/S13alsasound start' exits with status 0
: snd_cs46xxmodprobe: Can't locate module snd_cs46xx
failed
A check in modules.conf shows indeed that snd_cs46xx is being called, but the module is actually cs46xx. So I presume there is some invisible massaging, that has now broken. I did disable most sound modules, leaving only the generic ones, Soundblaster, Crystal, and OPL.
Anyone have a clue? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org
iEYEARECAAYFAj2ZxkwACgkQnQ18+PFcZJvVkgCbBOOjkCSnJVEczXnzZnaZVCdv lO4An2S1RC+yoKEQldURHLQG/Eb98L4o =8sLX -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 This is a deal-killer for us. Suse's recommendation is: http://sdb.suse.de/en/sdb/html/fhassel_vmware_segfault.html But I have 512MB. Nevertheless I tried it, and still a segfault on VM powerup. But only when I've turned off networking. With networking on, the whole network hangs when I start vmware daemons; power down is then the only option, as Shutdown is never able to complete. Googlers say 'ditch the Suse kernel'. Or at least recompile with Himem turned off. Well my compile went fine, but make modules crashes with some error that appears to be a coding problem. (compiling before entering the directory?) But I don't have the skills nor time to track this down: make[2]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers/ide' make -C ieee1394 modules_install make[2]: Entering directory `/usr/src/linux-2.4.18.SuSE/drivers/ieee1394' mkdir -p /lib/modules/2.4.18/kernel/drivers/ieee1394/ cp pcilynx.o ohci1394.o video1394.o raw1394.o sbp2.o dv1394.o eth1394.o amdtp.o cmp.o /lib/modules/2.4.18/kernel/drivers/ieee1394/ cp: cannot stat `pcilynx.o': No such file or directory cp: cannot stat `ohci1394.o': No such file or directory cp: cannot stat `video1394.o': No such file or directory cp: cannot stat `raw1394.o': No such file or directory cp: cannot stat `sbp2.o': No such file or directory cp: cannot stat `dv1394.o': No such file or directory cp: cannot stat `eth1394.o': No such file or directory cp: cannot stat `amdtp.o': No such file or directory cp: cannot stat `cmp.o': No such file or directory make[2]: *** [_modinst__] Error 1 make[2]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers/ieee1394' make[1]: *** [_modinst_ieee1394] Error 2 make[1]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers' make: *** [_modinst_drivers] Error 2 hydra:/usr/src/linux # rpm -qa|grep k_ docbook_3-3.1-199 docbook_4-4.1-200 k_deflt-2.4.18-154 hydra:/usr/src/linux # This is mission-critical for us. Will Suse answer? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2YYpwACgkQnQ18+PFcZJtmlwCcDxmisEaKzqeHCseNusYQvtoN /JcAn0KJhvuePhPNcgsiEoRhrnEhOdrK =BYS1 -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 This is a deal-killer for us. Suse's recommendation is: http://sdb.suse.de/en/sdb/html/fhassel_vmware_segfault.html But I have 512MB. Nevertheless I tried it, and still a segfault on VM powerup. But only when I've turned off networking. With networking on, the whole network hangs when I start vmware daemons; power down is then the only option, as Shutdown is never able to complete. Googlers say 'ditch the Suse kernel'. Or at least recompile with Himem turned off. Well my compile went fine, but make modules crashes with some error that appears to be a coding problem. (compiling before entering the directory?) But I don't have the skills nor time to track this down: make[2]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers/ide' make -C ieee1394 modules_install make[2]: Entering directory `/usr/src/linux-2.4.18.SuSE/drivers/ieee1394' mkdir -p /lib/modules/2.4.18/kernel/drivers/ieee1394/ cp pcilynx.o ohci1394.o video1394.o raw1394.o sbp2.o dv1394.o eth1394.o amdtp.o cmp.o /lib/modules/2.4.18/kernel/drivers/ieee1394/ cp: cannot stat `pcilynx.o': No such file or directory cp: cannot stat `ohci1394.o': No such file or directory cp: cannot stat `video1394.o': No such file or directory cp: cannot stat `raw1394.o': No such file or directory cp: cannot stat `sbp2.o': No such file or directory cp: cannot stat `dv1394.o': No such file or directory cp: cannot stat `eth1394.o': No such file or directory cp: cannot stat `amdtp.o': No such file or directory cp: cannot stat `cmp.o': No such file or directory make[2]: *** [_modinst__] Error 1 make[2]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers/ieee1394' make[1]: *** [_modinst_ieee1394] Error 2 make[1]: Leaving directory `/usr/src/linux-2.4.18.SuSE/drivers' make: *** [_modinst_drivers] Error 2 hydra:/usr/src/linux # rpm -qa|grep k_ docbook_3-3.1-199 docbook_4-4.1-200 k_deflt-2.4.18-154 hydra:/usr/src/linux # This is mission-critical for us. Will Suse answer? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2YY6kACgkQnQ18+PFcZJvZ9ACffnijmV3bOC+ImtAmfpc8zME8 WKgAnRJkN/eM87m8wvbX9Df8eTqbiA7I =8E6O -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 BTW, I posted this several times because KMail kept crashing on Send. And it didn't receive, so I didn't know they were gong out. Sorry. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj2Yb0gACgkQnQ18+PFcZJvnrwCfVBcsWmuebp9i/dOPcZz+PFn+ AKYAn1Xbk7HctZvb2zL6ng+k25kUOezD =cuFt -----END PGP SIGNATURE-----
participants (3)
-
Anders Johansson -
Carl -
Peter B Van Campen