Bug 1019415
Summary: | nm-online returns success before DHCP completes, causing NFS mounts to fail on boot | ||||||
---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Mike Chambers <mike> | ||||
Component: | NetworkManager | Assignee: | Dan Williams <dcbw> | ||||
Status: | CLOSED CURRENTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||
Severity: | unspecified | Docs Contact: | |||||
Priority: | unspecified | ||||||
Version: | 20 | CC: | bugzilla, dcbw, dr.diesel, eblake, johannbg, jonathan, lnykryn, msekleta, ngaywood, phil_g, plautrba, rdieter, samuel-rhbugs, systemd-maint, tchollingsworth, tflink, vpavlin, zbyszek | ||||
Target Milestone: | --- | ||||||
Target Release: | --- | ||||||
Hardware: | All | ||||||
OS: | Linux | ||||||
Whiteboard: | |||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2014-01-10 19:36:52 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: | |||||||
Attachments: |
|
Description
Mike Chambers
2013-10-15 16:37:17 UTC
Created attachment 812630 [details]
my /var/log/messages file
The system actually has auto mounted my nfs share *1* time, if you look at the boot sequence around the 11:38ish mark, you will see where it gets mounted, /home/download is the nfs directory. It looks like on this one, NetworkManager got 100% configured 1st, then maybe it was able to mount with network running?
All other times before it or after it have failed, and nothing was done before or after (as far as changes or commands) to make a difference. It seems on all the others, that NM isn't yet totally configured and therefore won't work?
After some debugging on test.o [1] we figured out that /usr/bin/nm-online seems to be returning before DHCP is fully configured. If you look at his logs, you'll notice that dhclient reports success after NetworkManager-wait-online.service successfully exits. To make double-sure, I had him add a extra service Before=network.target [2] that just waits until it can successfully grep his IP address from ifconfig, and his nfs mounts started to work again. Moving to NetworkManager since that ships NetworkManager-wait-online.service and /usr/bin/nm-online. [1] https://lists.fedoraproject.org/pipermail/test/2013-October/118240.html [2] https://lists.fedoraproject.org/pipermail/test/2013-October/118356.html Seconded! Thanks guys, the bug-report (and mail-discussion) really saved my christmas! (I changed my "Network Manager Wait Online" service description from using nm-online - which will just tell about NetworkManager being started or not, to ExecStart=/bin/bash -c "until ping -c1 [someIPonYourNetwork]; do sleep 1; continue; done" I added the 'sleep 1' because otherwise, it's putting many 'network unreachable' into the startup logs... ) NetworkManager-0.9.9.0-22.git20131003 (updates-testing) fixes the problem. (In reply to Steve from comment #4) This does indeed fix the problem, and everything seems to work as normal. Unless someone else has other issues with it, will close it if no issues. |