[opensuse-factory] Some strange hang at updating, check packages and Magic REISUB
Hi, I've been updating today to last snapshot and during that process everything stopped working. Browser (QupZilla) hanged, kwin marked it as hanged application, but nothing else happened after that. I couldn't switch to other window or desktop using hotkey or mouse. Mouse was still moving, but not doing anything else. I've tried to switch to tty1, which was successful, but after typing login root, no password prompt showed. Same for other tty and main user. Switching between tty's (text and gui) worked well, but I couldn't do anything else. HDD (SSD?) indicator was constantly lighted, as no any sound, I'd expect only ssd was used. So I've tried to wait, but as nothing happened, I've pressed ralt+sysrq+REISUB, REI is blocked for some reason by default in openSUSE, but SUB worked as expected, so I've rebooted. Last lines in `journalctl -b -1` are [1]. Last lines in `/var/log/zypp/ history` before reboot are [2]. I've tried continue updating, but got error [3]. `zypper clean` didn't help, after `rpm --rebuilddb` everything seems worked fine. I've reinstalled gcc7 to be sure it isn't screwed. So, as I haven't found any serious error in log, main question is how I can find out what really happened? I suspect some I/O problem, but not sure, and not sure what it was exactly. journalctl looks not helpful. Where else I can look for info? Second question is how I should check everything is ok with my packages? I've rebuilt rpm db, but are all files ok? Without reinstalling every package. And third question is why REI part isn't active. In such cases EI (TERM, KILL) look helpful, am I wrong? How I should activate them in right way (stable after reboot)? I've tried YaST (Security Center and Hardening), but there only on/off option (or maybe I misunderstood check/cross). Maybe it's somewhere in YaST's /etc/sysconfig, but I haven't found it (it good to have some search/ filter there). [1] `journalctl -b -1` Mar 07 11:05:35 thaumaturge [RPM][8818]: Transaction ID 5a9fab5f started Mar 07 11:05:35 thaumaturge [RPM][8818]: erase apparmor-utils- lang-2.12-3.2.noarch: success Mar 07 11:05:35 thaumaturge [RPM][8818]: install apparmor-utils- lang-2.12-4.1.noarch: success Mar 07 11:05:35 thaumaturge [RPM][8818]: erase apparmor-utils- lang-2.12-3.2.noarch: success Mar 07 11:05:35 thaumaturge [RPM][8818]: install apparmor-utils- lang-2.12-4.1.noarch: success Mar 07 11:05:35 thaumaturge [RPM][8818]: Transaction ID 5a9fab5f finished: 0 Mar 07 11:05:35 thaumaturge [RPM][8819]: Transaction ID 5a9fab5f started Mar 07 11:05:35 thaumaturge [RPM][8819]: erase gcc7-7.3.1+r258025-1.1.x86_64: success [2] `/var/log/zypp/history` 2018-03-07 11:05:34|install|PackageKit-backend-zypp|1.1.8-2.1|x86_64|| openSUSE-20170510-0|9b179af880eeb0cd7fd6e6721aa81abbf67fabf3| 2018-03-07 11:05:34|install|python3-requests|2.18.4-2.1|noarch|| openSUSE-20170510-0|a35fefd9d35f9b864cf83b1551bdeeb1eeb6d552| 2018-03-07 11:05:35|install|gtk2-lang|2.24.32-2.1|noarch||openSUSE-20170510-0| 799dcba4a2f8f975e9c8fa8bc8f6a26d182a6842| 2018-03-07 11:05:35|install|apparmor-utils-lang|2.12-4.1|noarch|| openSUSE-20170510-0|3d99e2664ebd63fc40a9331d001b990e8989d478| ^@^@^@^@.... a lot of this characters, in fact they are just zeros ...^@^@^@^@ [3] zypper (rpm) error: Installation of gtk2-immodule-amharic-2.24.32-2.1.x86_64 failed: Error: Subprocess failed. Error: RPM failed: error: db4 error(-30986) from dbcursor->c_get: DB_PAGE_NOTFOUND: Requested page not found error: error(-30986) getting "System/Libraries" records from Group index: DB_PAGE_NOTFOUND: Requested page not found error: gtk2-immodule-amharic-2.24.32-2.1.x86_64: install failed -- Kind regards, Mykola Krachkovsky -- Найкращі побажання, Микола Крачковський
On středa 7. března 2018 12:01:28 CET Mykola Krachkovsky wrote:
Second question is how I should check everything is ok with my packages? I've rebuilt rpm db, but are all files ok? Without reinstalling every package.
rpm -qaV This will verify files of every installed package. It takes a while. I recommend to redirect the output to a file for later examination. There will be lot of modified files even on healthy system (e.g. configuration files). Michal -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
середа, 7 березня 2018 р. 14:18:33 EET Michal Srb написано:
On středa 7. března 2018 12:01:28 CET Mykola Krachkovsky wrote:
Second question is how I should check everything is ok with my packages? I've rebuilt rpm db, but are all files ok? Without reinstalling every package.
rpm -qaV
This will verify files of every installed package. It takes a while. I recommend to redirect the output to a file for later examination. There will be lot of modified files even on healthy system (e.g. configuration files).
Michal
Thanks, It looks almost ok to me, full at [1] Just several /usr files are suspicious: ....L.... /usr/share/YaST2/theme/current/icons (looks ok to me, link) .......T. /usr/lib64/gconv/gconv-modules.cache (idk) ....L.... d /usr/share/man/man1/ftp.1.gz (link to alternative, ok?) .M....... d /usr/share/man/man1/npm.1.gz (ok?) missing /usr/bin/ecpg missing /usr/bin/pg_config (both are from installed postgresql96-devel, which was installed/updated couple weeks ago: 2018-02-13 12:11:48|install|postgresql96-devel|9.6.7-1.1|x86_64|| openSUSE-20170510-0|3e09985a5dc3359bfded1dc4c1e3f86642117279|, I'll reinstall it later) .M....... d /usr/share/man/man1/node.1.gz (ok?) ....L.... /usr/share/java/xml-commons-apis.jar (link to alternative, ok?) [1] http://susepaste.org/63679604 -- Kind regards, Mykola Krachkovsky -- Найкращі побажання, Микола Крачковський
On 03/07/2018 03:01 AM, Mykola Krachkovsky wrote:
And third question is why REI part isn't active. In such cases EI (TERM, KILL) look helpful, am I wrong?
For safety reasons, obviously, Alt-SysRq is defaulted with some actions disconnected.
How I should activate them in right way (stable after reboot)? I've tried YaST (Security Center and Hardening), but there only on/off option (or maybe I misunderstood check/cross). Maybe it's somewhere in YaST's /etc/sysconfig, but I haven't found it (it good to have some search/ filter there).
Add the line: kernel.sysrq = 1 in /etc/sysctl.conf -- -Gerry Makaro openSUSE Member openSUSE Forum Moderator openSUSE Contributor aka Fraser_Bell on the Forums, OBS, IRC, and mail at openSUSE.org Fraser-Bell on Github -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
середа, 7 березня 2018 р. 23:55:21 EET Fraser_Bell написано:
On 03/07/2018 03:01 AM, Mykola Krachkovsky wrote:
And third question is why REI part isn't active. In such cases EI (TERM, KILL) look helpful, am I wrong?
For safety reasons, obviously, Alt-SysRq is defaulted with some actions disconnected.
Yes, I expected something like that, just haven't understand why EI are blocked while UB aren't.
Add the line:
kernel.sysrq = 1
in /etc/sysctl.conf
Thanks! I've just set 248 instead of default 184, that sounds enough for me. -- Kind regards, Mykola Krachkovsky -- Найкращі побажання, Микола Крачковський
Hi, I've got this problem second time yesterday. Again while installing updates. journalctl again looks ok to me. smartctl also has no errors. And this hanging haven't happened before or while other actions are done. I've pressed Ctrl+C when zypper hanged (progress stopped dead), it said something about "exiting gracefully", but didn't exit actually, though I've pressed ^C a lot. It really looks like some I/O related hang/deadlock in the kernel. I don't suspect hardware much as it hadn't happened before or in other situations, though statistics isn't rich atm. Any hints? FS: ext4, ssd with enabled discard. -- Kind regards, Mykola Krachkovsky -- Найкращі побажання, Микола Крачковський
Any problem appearing on your logs? Currently there is this kernel bug that affects some of us: https://bugzilla.kernel.org/show_bug.cgi?id=198861 see also here: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=891467 It also has been discussed in this list, or rather not discussed. https://lists.opensuse.org/opensuse-factory/2018-02/msg01130.html -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
субота, 10 березня 2018 р. 21:18:25 EET Christos Gourdoupis написано:
Any problem appearing on your logs? Problem is I can't find anything really bad, [1] is final part of log.
Currently there is this kernel bug that affects some of us: https://bugzilla.kernel.org/show_bug.cgi?id=198861 see also here: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=891467
It also has been discussed in this list, or rather not discussed. https://lists.opensuse.org/opensuse-factory/2018-02/msg01130.html
Thanks, but I can't find similar entries in my log, all error|exception|fail in [2]. All ata [3], only strange is "supports DRM functions and may not be fully accessible". PS full log [4], maybe I'm missing something important. [1] http://susepaste.org/19421536 [2] http://susepaste.org/69618477 [3] http://susepaste.org/91708947 [4] https://drive.google.com/open?id=11oZ01DS0GHF-19lJw7aU5xvWuRM26IqU -- Kind regards, Mykola Krachkovsky -- Найкращі побажання, Микола Крачковський
participants (4)
-
Christos Gourdoupis -
Fraser_Bell -
Michal Srb -
Mykola Krachkovsky