Summary: | doesn't restore the DHCP configuration after re-plugging | ||
---|---|---|---|
Product: | Branch 4.1 | Reporter: | Ivan Zakharyaschev <imz> |
Component: | etcnet | Assignee: | Nobody's working on this, feel free to take it <nobody> |
Status: | CLOSED WONTFIX | QA Contact: | qa-4.1 <qa-4.1> |
Severity: | normal | ||
Priority: | P2 | CC: | gray_graff, mike, pilot, sem |
Version: | unspecified | ||
Hardware: | all | ||
OS: | Linux |
Description
Ivan Zakharyaschev
2008-12-28 19:59:06 MSK
Reproduceable. Forgot to say: it is ifplugd-0.28-alt4 which is in this distro. BTW, this problem doesn't occur in a 4.0 system (Lite 4.0.x): $ rpm -q ifplugd etcnet dhcpcd ifplugd-0.28-alt3 etcnet-0.9.7-alt0.M40.1 dhcpcd-1.3.22pl4-alt3.0 $ but, as I have described in the report, occurs in a 4.1 system (Desktop 4.1): $ rpm -q ifplugd etcnet dhcpcd ifplugd-0.28-alt4 etcnet-0.9.7-alt0.M41.1 dhcpcd-3.0.17-alt4 $ (In reply to comment #2) > BTW, this problem doesn't occur in a 4.0 system (Lite 4.0.x): So, the only significant difference seems to be in dhcpcd. > $ rpm -q ifplugd etcnet dhcpcd > ifplugd-0.28-alt3 > etcnet-0.9.7-alt0.M40.1 > dhcpcd-1.3.22pl4-alt3.0 > $ > > but, as I have described in the report, occurs in a 4.1 system (Desktop 4.1): > > $ rpm -q ifplugd etcnet dhcpcd > ifplugd-0.28-alt4 > etcnet-0.9.7-alt0.M41.1 > dhcpcd-3.0.17-alt4 > $ (In reply to comment #3) > (In reply to comment #2) > > BTW, this problem doesn't occur in a 4.0 system (Lite 4.0.x): > > So, the only significant difference seems to be in dhcpcd. The problem persists with dhcpcd-4.0.15-alt2 from Sisyphus. > > but, as I have described in the report, occurs in a 4.1 system (Desktop 4.1): > > > > $ rpm -q ifplugd etcnet dhcpcd > > ifplugd-0.28-alt4 > > etcnet-0.9.7-alt0.M41.1 > > dhcpcd-3.0.17-alt4 > > $ (In reply to comment #0) > ... > > Good. Then it was unplugged: > > Dec 28 17:20:39 localhost NetworkManager: <info> (eth0): carrier now OFF > (device state 1) > Dec 28 17:20:40 localhost ifplugd(eth0)[4829]: Link beat lost. > Dec 28 17:20:51 localhost ifplugd(eth0)[4829]: Executing > '/etc/net/scripts/ifplugd.action eth0 down'. > ... > > Then it was plugged back: > > Dec 28 19:26:07 localhost kernel: sky2 eth0: Link is up at 100 Mbps, full > duplex, flow control both > Dec 28 19:26:07 localhost NetworkManager: <info> (eth0): carrier now ON > (device state 1) > Dec 28 19:26:08 localhost ifplugd(eth0)[4829]: Link beat detected. > Dec 28 19:26:09 localhost ifplugd(eth0)[4829]: Executing > '/etc/net/scripts/ifplugd.action eth0 up'. > Dec 28 19:26:09 localhost dhcpcd[19510]: eth0: dhcpcd already running > (/var/run/dhcpcd-eth0.pid) > Dec 28 19:26:09 localhost ifplugd(eth0)[4829]: client: ..Error, eth0: dhcpcd > already running (/var/run/dhcpcd-eth0.pid) So, an even more probable reason for the problem is that the used iflugd (ifplugd-0.28-alt4) doesn't know how to talk correctly to the newer dhcpcd: there is no such error in the 4.0 system with a dhcpcd-1.3.22pl4-alt3.0. Perhaps, the newer dhcpcd (> 3.0) should conflict with this version of ifplugd, so that dhcpcd wouldn't be upgraded in a system with the old incompatible ifplugd. And ifplugd should be fixed. (In reply to comment #5) > > Dec 28 19:26:09 localhost ifplugd(eth0)[4829]: Executing > > '/etc/net/scripts/ifplugd.action eth0 up'. > > Dec 28 19:26:09 localhost dhcpcd[19510]: eth0: dhcpcd already running > > (/var/run/dhcpcd-eth0.pid) > > Dec 28 19:26:09 localhost ifplugd(eth0)[4829]: client: ..Error, eth0: dhcpcd > > already running (/var/run/dhcpcd-eth0.pid) > > So, an even more probable reason for the problem is that the used iflugd Sorry, it's clear that the problem is in the script provided by etcnet: $ rpm -qf /etc/net/scripts/ifplugd.action etcnet-0.9.7-alt0.M41.1 $ > (ifplugd-0.28-alt4) doesn't know how to talk correctly to the newer dhcpcd: > there is no such error in the 4.0 system with a dhcpcd-1.3.22pl4-alt3.0. > > Perhaps, the newer dhcpcd (> 3.0) should conflict with this version of ifplugd, > so that dhcpcd wouldn't be upgraded in a system with the old incompatible > ifplugd. > > And ifplugd should be fixed. Perhaps, the solution is the solution of https://bugzilla.altlinux.org/show_bug.cgi?id=19582 . Sorry, I'm largely out of that context by now. Just in case, look at ifplugstatus(8) to verify that there's no additional problem with e.g. kernel driver or ifplugd link detection. could you test recent etcnet please ? В 4.1/branch исправления не будут вноситься уже технически (заглушена очередь на сборку), поэтому прошу ошибки, актуальные для sisyphus/p7/t7, перевесить на текущие ветки или сизиф. |