From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020529 Description of problem: After configuring a new modem connection (USB modem on /dev/input/ttyACM0), attempting to activate it gives "Cannot activate newtork device <provider_name>! Failed to activate ppp0 with error 28". There is no user-level documentation on this error, not even via google or google/groups. Version-Release number of selected component (if applicable): redhat-config-network-1.2.13-1 How reproducible: Always Steps to Reproduce: 1.Configure USB modem connection: /dev/input/ttyACM0 [ignore warning of "No modem was found on your system"], 115200 baud, CRTSCTS flow control; enter phone number, provider, login, password, ask for auto IP and auto DNS; save. 2.Attempt to activate: fails. 3.Reboot; attempt to activate: still fails. Actual Results: "Cannot activate network device <provider_name>! Failed to activate ppp0 with error 28". Expected Results: Modem should dial provider, and activate network upon connection. Additional info: Partial workaround: su, then manually invoke "wvdial <provider_name>". Modem dials phone, connection is established, wvdial invokes pppd. But there is no DNS, so you must use numerical addresses.
does a # usernetctl <provider_name> up work? or a # /sbin/ifup <provider_name> ?? both as root..
the initscripts call wvdial like this: /usr/bin/wvdial --remotename <providername> <providername> and later on pppd calls: /etc/ppp/ip-up which then uses $LOGDEVICE (which is the remotename) to determine the config script. Then ifup-post does the DNS setting... in your case $LOGDEVICE is not set to providername, thus the setting fails. you can either rename the <providername> device to ppp0 by changing its Nickname in redhat-config-network or we debug, why ifup/usrnetctl does not work.
*** Bug 98786 has been marked as a duplicate of this bug. ***
Created attachment 92835 [details] strace -f usernetctl <provider> up Done while logged-in as root.
Created attachment 92836 [details] strace -f ifup <provider> Done while logged-in as root.
Created attachment 92837 [details] ls -l of files where EACCES = access(file, X_OK) /etc/bashrc, /etc/sysconfig/init, /etc/sysconfig/i18n are present and readable by anybody, but do not have execute permission, hence testing for X_OK fails, although the shell will still execute them.
err... # sh -x ifup <provider> would be more readable :)
or # sh -x /sbin/ifup <provider>
Created attachment 92838 [details] cd /etc; find . -name '*<provider>*' All the files in /etc or below with <provider> in the filename.
Created attachment 92839 [details] contents of files with <provider> in filename All the ifcfg-<provider> files are hard-linked to each other.
Created attachment 92840 [details] sh -x /sbin/ifup <provider> Still fails: + exec /etc/sysconfig/network-scripts/ifup-ppp ../networking/profiles/default/ifcfg-Olympus Failed to activate ppp0 with error 28
# sh -x /etc/sysconfig/network-scripts/ifup-ppp ../networking/profiles/default/ifcfg-Olympus
Created attachment 92843 [details] sh -x /etc/sysconfig/network-scripts/ifup-ppp /etc/sysconfig/networking/profiles/default/ifcfg-Olympus Still fails: + exec /sbin/ppp-watch ppp0 '' Failed to activate ppp0 with error 28
The suggested workaround: "wvdial --remotename <provider> <provider>" FAILS to set DNS. [Note <provider> appears twice, once for --remotename, and once more.] I still have to copy the DNS server settings manually from /var/log/messages to the Network Configuration > DNS tab, then File > Save, and then DNS works.
*** This bug has been marked as a duplicate of 97845 ***
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.