[opensuse-factory] TW 32bit net install fails: device=eth0 failed
Is there any way to troubleshoot this? tty2 hasn't yet provided a shell at this failure point. Known problem? Tried a second time: same problem. Among messages on tty3: setup_interface: eth0 already up error 6: could not resolve host: download.opensuse.org (or ftp5.gwdg.de) no openSUSE Tumbleweed repository found device not found (err = 1): eth0 initrd takes ~11.5 minutes for Grub to load with 2.0GHz CPU from PATA HD. -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 19 Sep 2018 03:43:29 -0400 Felix Miata wrote:
Is there any way to troubleshoot this? tty2 hasn't yet provided a shell at this failure point. Known problem? Tried a second time: same problem. I also tried to install the net install on i586 HW recently and observed similar problems.
There is a console on tty9. When I try to ping <IP addr> there I get an error message like fips.c:508: OpenSSL internal error, FATAL FIPS SELFTEST FAILURE Pinging the machine from another machine works, so the network itself is operational. I am not using DHCP but entered the network configuration manually. I have the impression that any outgoing traffic fails due to this FIPS error when even ping is not working. Trying the i586 net install iso in a VM on 64bit HW works. So I suspect the fips selftest fails on real i586 HW. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 19 Sep 2018 10:02:18 +0200 dieter wrote:
When I try to ping <IP addr> there I get an error message like fips.c:508: OpenSSL internal error, FATAL FIPS SELFTEST FAILURE
...
Trying the i586 net install iso in a VM on 64bit HW works. So I suspect the fips selftest fails on real i586 HW. I did some more testing in the VM and I found out: in the VM the CPU had the "sse2" feature which my 32-bit HW does not have.
Removing the sse2 feature from the VM reproduces the error message when trying ping: crypto/fips/fips.c:508: OpenSSL internal error: FATAL FIPS SELFTEST FAILURE and no repository is accessible. So the 32-bit net install requires a sse2 capable CPU which is the problem in my case. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
dieter composed on 2018-09-19 12:13 (UTC+0200):
So the 32-bit net install requires a sse2 capable CPU which is the problem in my case.
Apparently there's more requiring sse2 than just the net installer. The reason I was trying to install was an existing TW installation was core dumping trying to zypper dup, so I wanted to try a fresh install. I first Googled for release notes and system requirements looking for any mention of sse2 without finding any before proceeding to try to install. I also tried to install from openSUSE-Tumbleweed-DVD-i586-Current.iso. It failed with exit code 134. As of Fedora 28, sse2 has not become a requirement yet, but I guess it's time to abandon openSUSE for Debian on this box and any others with Athlon class CPUs. Is there any easy way to see if a CPU provides sse2, something better than detecting its absence from /proc/cpuinfo's flags? Has anyone found mention of sse2 in any installation doc? -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Felix Miata
Is there any easy way to see if a CPU provides sse2, something better than detecting its absence from /proc/cpuinfo's flags? Has anyone found mention of sse2 in any installation doc?
inxi -Cf |grep sse2 -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 19 Sep 2018 12:36:45 -0400 Felix Miata wrote:
I also tried to install from openSUSE-Tumbleweed-DVD-i586-Current.iso. It failed with exit code 134.
As of Fedora 28, sse2 has not become a requirement yet, but I guess it's time to abandon openSUSE for Debian on this box and any others with Athlon class CPUs.
Is there any easy way to see if a CPU provides sse2 According to wikipedia only 64-bit AMD CPUs support sse2, and Intel only from Pentium 4 up. For me it's a slightly strange choice to build the 32-bit distribution for this very narrow selection of HW - and completely inaccurate to name the media "i586" when it wont work on most i686 CPUs. Probably it is mainly targeted for VMs and not real HW. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 19.09.18 um 18:36 schrieb Felix Miata:
As of Fedora 28, sse2 has not become a requirement yet, but I guess it's time to abandon openSUSE for Debian on this box and any others with Athlon class CPUs.
...or file a bug report about it? Maybe this is an unintended side effect of some change that just nobody reported yet? -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Stefan Seyfried composed on 2018-09-19 21:52 (UTC+0200):
Felix Miata composed:
As of Fedora 28, sse2 has not become a requirement yet, but I guess it's time to abandon openSUSE for Debian on this box and any others with Athlon class CPUs.
...or file a bug report about it?
Maybe this is an unintended side effect of some change that just nobody reported yet?
Apparently it did nearly 18 months ago, and got wontfixed: https://bugzilla.opensuse.org/show_bug.cgi?id=1032165 -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 19 Sep 2018 16:25:34 -0400 Felix Miata wrote:
...or file a bug report about it?
Maybe this is an unintended side effect of some change that just nobody reported yet?
Apparently it did nearly 18 months ago, and got wontfixed: https://bugzilla.opensuse.org/show_bug.cgi?id=1032165 Thanks, the information in this bug report was extremely helpful. I managed to install the current Tumbleweed snapshot (20180917) in my VM without SSE2 support using the net install iso.
Easy check of the SSE2 support: grep sse2 /proc/cpuinfo This is what I did it in case it's useful to somebody, use the information at your own risk - and be aware the outcome might change with each new snapshot. 1 boot the net install iso 2 when the error message appears that no repo is accessible *AND* the reason is the lack of SSE2 support of your CPU: switch to tty9, verify that pinging a host which should be accessible, e.g. your gateway, results in an error message like "crypto/fips/fips.c:508: OpenSSL internal error: FATAL FIPS SELFTEST FAILURE" then remove the hmac files from /usr/lib (this affects only the installation system booted from the iso, it is not necessary to meddle with the initrd of the installation media) rm /usr/lib/.lib*hmac Verify that "ping" now works, e.g. ping your gateway 3 switch back to tty1 4 in the installer menu go back and enter the network settings, it's not necessary to change any settings, but it seems to be needed in order to make the installer retry the access to the repos 5 proceed with the installation as usual and have a lot of fun. There are some applications missing in the official 32-bit Tumbleweed repos, e.g. Firefox (seamonkey is available, so it's not a problem). With this result I am seriously considering to update my openSUSE 13.2 32-bit system to Tumbleweed. BTW: the (current) i686 rescue image of Tumbleweed starts without problems and has network connectivity on a CPU without SSE2 support, it even contains Firefox 61 which also works. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
dieter composed on 2018-09-20 12:58 (UTC+0200):
5 proceed with the installation as usual and have a lot of fun.
What happens on your completed installation when you try to do 'zypper dup'? Here it exploded at around half done, which is what led to me trying to do a fresh TW installation. -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Fri, 21 Sep 2018 03:41:23 -0400 Felix Miata wrote:
What happens on your completed installation when you try to do 'zypper dup'? Here it exploded at around half done, which is what led to me trying to do a fresh TW installation.
The original installation described in my last mail was with snapshot 20180917. The zypper dup to snapshot 20180919 completed without problems and also no issues were observed after reboot. My experiments are currently still using a VM with deactivated SSE2 feature, on real HW problems could arise from no-longer supported devices. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
dieter composed on 2018-09-21 12:26 (UTC+0200):
On Fri, 21 Sep 2018 03:41:23 -0400 Felix Miata wrote:
What happens on your completed installation when you try to do 'zypper dup'? Here it exploded at around half done, which is what led to me trying to do a fresh TW installation.
The original installation described in my last mail was with snapshot 20180917. The zypper dup to snapshot 20180919 completed without problems and also no issues were observed after reboot. My experiments are currently still using a VM with deactivated SSE2 feature, on real HW problems could arise from no-longer supported devices.
There's the difference, simulated testing. The only VMs I ever use are for DOS. I test Linux only on real hardware. -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (4)
-
dieter -
Felix Miata -
Patrick Shanahan -
Stefan Seyfried