Sorry to chooce the dhcpcd component because pump is not there. I use pump to connect to my internet ISP. I use a cable-modem. I use a network card connected in the slot 0 of the pcmcia connector. I receive well my IP address and all the other address (DNS,...). One thing important : the lease time is very short; only 20 minutes. Until the lease time is finished, I can do the following command without problem : pump --status. If the lease time is finished (in fact the DHCP server (witch is NT I thing, because, I had to install the new pump package)), this command doesn't answer. A CTRL-C and I can goes futher. If I forgot to do that blocking command and I do a shutdown of Linux, I can't do a CTRL-C (normal, I'm shutting down!), I have no other choice to power off the machine : bad isn't it! By the way, the Internet is still accessible, I can use netscape without problem. Pierre Thonne Brussels
*** Bug 3296 has been marked as a duplicate of this bug. *** I've got RH6 working just fine, except for one problem - "pump". I've installed the latest version from updates: [root@cs2867-105 /root]# rpm -q pump pump-0.6.7-1 and it assigns an IP address just fine via DHCP at bootup (I'm an Austin RoadRunner cablemodem subscriber): [root@cs2867-105 /root]# ifconfig eth0 eth0 Link encap:Ethernet HWaddr 00:A0:CC:3E:E0:71 inet addr:24.28.67.225 Bcast:24.28.67.255 Mask:255.255.255.0 UP BROADCAST RUNNING MTU:1500 Metric:1 RX packets:55484 errors:0 dropped:0 overruns:0 frame:0 TX packets:51350 errors:0 dropped:0 overruns:0 carrier:0 collisions:1128 txqueuelen:100 Interrupt:11 Base address:0xe800 However, it seems that the dhcp lease is *not* being renewed - forcing me to stop/restart the network interface every few hours, in order to either re-grab the same IP address, or be assigned a new one. Other strangeness: [root@cs2867-105 /root]# pump -i eth0 -s Operation failed. [root@cs2867-105 /root]# pump -R Operation failed. [root@cs2867-105 /root]# pump -s Operation failed. System is an Intel Celeron 366, Abit BH6 motherboard, 17.2gig IDE HD, Netgear FA310TX (DEC Tulip chipset 10/100bt) ethernet card, Riva TNT (AGP) video, PS/2 keyboard and mouse. 256mb RAM. ------- Additional Comments From mrbill 06/06/99 03:33 ------- My behavior is similar to that experienced in bug #3227, btw.
Have you tried the new pump from the errata? Lease renewal was definitely broken... Please let us know if that doesn't fix these problems.
I'm using the pump package from the errata. I had to use it because with the original one, i cannot connect to my ISP (cable-modem). So i was using the pump-0.6.7-1.i386.rpm package when I first describe the problem here. Cheers, Pierre ------- Additional Comments From 06/23/99 05:53 ------- This *is* the version from the errata, unfortunately; I applied all relevant updates immediately after I installed the OS. -bill (mrbill)
I've been told that the newest version of pump fixes these problems. (ftp://ftp.redhat.com/rawhide/i386/RedHat/RPMS/) This is version 0.7.3-2 or something like that. Beware that this is a "bleeding edge" release and as such it may not be terribly stable. Anyway, I was having the exact same problem with pump. At first I thought it was a problem with IP Masquerading (since I'm using the box as a proxy server). It took me several days of having the machine work fine for an hour, then not connecting to the internet at all. I couldn't ping any public IP addresses from either the NATed hosts or the linux box itself. When I rebooted it worked fine for about another hour. I finally caught on when I realized that my cable modem company's DHCP server sets a lease time of (you guessed it) one hour. Pump was completely locking up when it tried to renew the DHCP address, taking the entire eth0 interface down with it. If I killed it (using a killall pump command) and then brought it back up it worked fine until it had to renew again. Also noticed that "pump -R" or "pump -R -i eth0" never worked even after I upgraded pump to 0.7.0. Anyway, to make a long story short (too late, I know) I just did away with pump and went back to good 'ol dhcpcd. It works perfectly for my purposes and it hasn't gone down since.
There will be a new pump (0.7.6) on ftp://people.redhat.com/ewt this afternoon, which I believe fixes these problems. If it does not help, please reopen this bug.