Bug 994268
| Summary: | enabling autofs causes systemd ordering cycle | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Ray Mikkelson <ray.mikkelson> |
| Component: | systemd | Assignee: | systemd-maint |
| Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | urgent | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 18 | CC: | johannbg, lnykryn, maurizio.antillon, msekleta, notting, plautrba, ray.mikkelson, systemd-maint, vpavlin, zbyszek |
| Target Milestone: | --- | Keywords: | Reopened |
| Target Release: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | systemd-201-2.fc18.9 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2013-11-13 02:25:00 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Ray Mikkelson
2013-08-06 23:00:43 UTC
How is software-linux.mount configured? Via fstab? Can you provide the precise line? Lennart, No, /software/linux is automounted via NIS. So, to walk down the tree.... /etc/auto.master has (I'm only including the pertinent lines for clarity): /software yp:auto_software --timeout=0 Doing a ypcat -k auto_software shows: Linux -rw,hard,retry=0,noatime,nodiratime pie:/vol0/software/linux linux -rw,hard,retry=0,noatime,nodiratime pie:/vol0/software/linux (I'm implementing an "alias" for Linux to "point to" linux) My /etc/fstab has no entries related to /software/linux at all. The output from mount | grep -i linux shows: pie:/vol0/software/linux on /software/Linux type nfs4 (rw,noatime,nodiratime,vers=4.0,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=192.168.1.164,local_lock=none,addr=192.168.1.60) pie:/vol0/software/linux on /software/linux type nfs4 (rw,noatime,nodiratime,vers=4.0,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=192.168.1.164,local_lock=none,addr=192.168.1.60) Note that this all worked just fine in F16... This should be fixed by http://cgit.freedesktop.org/systemd/systemd/commit/?id=7700945. I guess we need to backport this here too. systemd-204-17.fc19 has been submitted as an update for Fedora 19. https://admin.fedoraproject.org/updates/systemd-204-17.fc19 Any chance that the fix will be ported back to F18? It should be. I'll try to produce an update, but because of the version gap, backporting stuff isn't easy. Package systemd-204-17.fc19: * should fix your issue, * was pushed to the Fedora 19 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing systemd-204-17.fc19' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-2013-19730/systemd-204-17.fc19 then log in and leave karma (feedback). systemd-204-17.fc19 has been pushed to the Fedora 19 stable repository. If problems still persist, please make note of it in this bug report. systemd-201-2.fc18.9 has been submitted as an update for Fedora 18. https://admin.fedoraproject.org/updates/systemd-201-2.fc18.9 Package systemd-201-2.fc18.9: * should fix your issue, * was pushed to the Fedora 18 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing systemd-201-2.fc18.9' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-2013-20267/systemd-201-2.fc18.9 then log in and leave karma (feedback). Tested with: systemd.x86_64 201-2.fc18.9 @updates-testing (Thanks for porting the fix back!!) Fix does solve the ordering cycle problem, which is a big improvement. However, being admittedly somewhat naive here, shouldn't the shutdown messages in /var/log/messages be the same each time? I'm guessing that there might be timing issues with syslogd... Is there a reliable place to see all of the messages output during shutdown? What is stored in the messages file doesn't match what I see on the screen (momentarily) while the system shuts down. Anyway, to document the testing that I did, here's the pertinent output from /var/log/messages, with my comments: # # Previous shutdown without fix # Oct 28 14:17:57 security systemd-logind[14668]: System is powering down. Oct 28 14:17:57 security systemd[1]: Found ordering cycle on network.target/stop Oct 28 14:17:57 security systemd[1]: Walked on cycle path to NetworkManager.service/stop Oct 28 14:17:57 security systemd[1]: Walked on cycle path to dbus.socket/stop Oct 28 14:17:57 security systemd[1]: Walked on cycle path to sysinit.target/stop Oct 28 14:17:57 security systemd[1]: Walked on cycle path to local-fs.target/stop Oct 28 14:17:57 security systemd[1]: Walked on cycle path to software-FedoraMirror.mount/stop Oct 28 14:17:57 security systemd[1]: Walked on cycle path to network.target/stop Oct 28 14:17:57 security systemd[1]: Breaking ordering cycle by deleting job NetworkManager.service/stop Oct 28 14:17:57 security systemd[1]: Job NetworkManager.service/stop deleted to break ordering cycle starting with network.target/stop Oct 28 14:17:57 security systemd[1]: Starting Show Plymouth Power Off Screen... Oct 28 14:17:57 security systemd[1]: Deactivating swap /dev/dm-0... Oct 28 14:17:57 security systemd[1]: Deactivating swap /dev/dm-0... Oct 28 14:17:57 security systemd[1]: Deactivating swap /dev/dm-0... Oct 28 14:17:57 security systemd[1]: Deactivating swap /dev/dm-0... Oct 28 14:17:57 security systemd[1]: Deactivating swap /dev/dm-0... Oct 28 14:17:57 security systemd[1]: Stopping Sound Card. Oct 28 14:17:57 security systemd[1]: Stopping Sockets. Oct 28 14:17:57 security rsyslogd: [origin software="rsyslogd" swVersion="7.2.6" x-pid="606" x-info="http://www.rsyslog.com"] exiting on signal 15. # # First shutdown immediately after installing new systemd packages # NFS directories are automount'ed # Oct 31 13:11:35 security systemd-logind[26978]: System is powering down. Oct 31 13:11:35 security systemd[1]: Starting Show Plymouth Power Off Screen... Oct 31 13:11:35 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:11:35 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:11:35 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:11:35 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:11:35 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:11:35 security systemd[1]: Stopping Sound Card. Oct 31 13:11:35 security systemd[1]: Stopped target Sound Card. ct 31 13:11:35 security systemd[1]: Stopping Import network configuration from initramfs... Oct 31 13:11:35 security systemd[1]: Stopped Import network configuration from initramfs. Oct 31 13:11:35 security rsyslogd: [origin software="rsyslogd" swVersion="7.2.6" x-pid="731" x-info="http://www.rsyslog.com"] exiting on signal 15. # # 2nd shutdown after install, no NFS directories mounted # Oct 31 13:13:24 security systemd-logind[667]: System is powering down. Oct 31 13:13:24 security systemd[1]: Starting Show Plymouth Power Off Screen... Oct 31 13:13:24 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:13:24 security rsyslogd: [origin software="rsyslogd" swVersion="7.2.6" x-pid="726" x-info="http://www.rsyslog.com"] exiting on signal 15. # # 3rd shutdown after install, NFS directories are automount'ed # Oct 31 13:16:15 security systemd-logind[666]: System is powering down. Oct 31 13:16:15 security systemd[1]: Starting Show Plymouth Power Off Screen... Oct 31 13:16:15 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:16:15 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:16:15 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:16:15 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:16:15 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:16:15 security systemd[1]: Stopping Sound Card. Oct 31 13:16:15 security systemd[1]: Stopped target Sound Card. Oct 31 13:16:15 security systemd[1]: Stopping Import network configuration from initramfs... Oct 31 13:16:15 security systemd[1]: Stopped Import network configuration from initramfs. Oct 31 13:16:15 security systemd[1]: Stopping Stop Read-Ahead Data Collection 10s After Completed Startup. Oct 31 13:16:15 security systemd[1]: Stopped Stop Read-Ahead Data Collection 10s After Completed Startup. Oct 31 13:16:15 security systemd[1]: Stopping CUPS Printing Service... Oct 31 13:16:15 security systemd[1]: Stopping Graphical Interface. Oct 31 13:16:15 security systemd[1]: Stopped target Graphical Interface. Oct 31 13:16:15 security rsyslogd: [origin software="rsyslogd" swVersion="7.2.6" x-pid="677" x-info="http://www.rsyslog.com"] exiting on signal 15. # # 4th shutdown, NFS directories automount'ed # Oct 31 13:23:22 security systemd-logind[674]: System is powering down. Oct 31 13:23:22 security systemd[1]: Starting Show Plymouth Power Off Screen... Oct 31 13:23:22 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:23:22 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:23:22 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:23:22 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:23:22 security systemd[1]: Deactivating swap /dev/dm-0... Oct 31 13:23:22 security systemd[1]: Stopping Sound Card. Oct 31 13:23:22 security systemd[1]: Stopped target Sound Card. Oct 31 13:23:22 security systemd[1]: Stopping Import network configuration from initramfs... Oct 31 13:23:22 security systemd[1]: Stopped Import network configuration from initramfs. Oct 31 13:23:22 security systemd[1]: Stopping Stop Read-Ahead Data Collection 10s After Completed Startup. Oct 31 13:23:22 security systemd[1]: Stopped Stop Read-Ahead Data Collection 10s After Completed Startup. Oct 31 13:23:22 security systemd[1]: Stopping CUPS Printing Service... Oct 31 13:23:22 security systemd[1]: Stopping Graphical Interface. Oct 31 13:23:22 security systemd[1]: Stopped target Graphical Interface. Oct 31 13:23:22 security systemd[1]: Stopping Restorecon maintaining path file context... Oct 31 13:23:22 security systemd[1]: Stopping ACPI Event Daemon... Oct 31 13:23:22 security systemd[1]: Stopping Console Mouse manager... Oct 31 13:23:22 security systemd[1]: Stopping Virtualization daemon... Oct 31 13:23:22 security systemd[1]: Stopping Kernel Samepage Merging (KSM) Tuning Daemon... Oct 31 13:23:22 security rsyslogd: [origin software="rsyslogd" swVersion="7.2.6" x-pid="682" x-info="http://www.rsyslog.com"] exiting on signal 15. (In reply to Ray Mikkelson from comment #11) > Tested with: > systemd.x86_64 201-2.fc18.9 > @updates-testing > > (Thanks for porting the fix back!!) > > Fix does solve the ordering cycle problem, which is a big improvement. Great. > However, being admittedly somewhat naive here, shouldn't the shutdown > messages in /var/log/messages be the same each time? I'm guessing that > there might be timing issues with syslogd... Is there a reliable place to > see all of the messages output during shutdown? What is stored in the > messages file doesn't match what I see on the screen (momentarily) while the > system shuts down. No, messages can be lost. The problem is that because of a kernel limitation, the message queue length cannot be changed, and is on 10 packets long. This means that at times when many messages are logged (like shutdown), journald cannot keep up with the flow. We can either lose messages, or make the message passing blocking. In later versions, it is blocking, but the downside is that it slows things down, and can even dead-lock if important stuff is waiting for logging, so special care must be taken. And the same thing happens on the interface between journald and syslog. If syslog is slow for whatever reason, it'll lose messages. You might have better luck with journalctl output. systemd-201-2.fc18.9 has been pushed to the Fedora 18 stable repository. If problems still persist, please make note of it in this bug report. |