Re: [Dnsmasq-discuss] mixup of tftp-root and boot filename
Carlos Carvalho wrote: Simon Kelley (si...@thekelleys.org.uk) wrote on 3 August 2008 21:02: Carlos Carvalho wrote: I have tftp-root=/var/remoteboot in dnsmasq.conf. For a machine I have in dhcp-options: ometepe,vl5,67,/tftpboot/pxelinux.0 In the log there is dnsmasq[16090]: sent size: 21 option: 67:bootfile-name 2f:74:66:74:70:62:6f:6f:74:2f:70:78:65... Checking with the ascii table this looks correct. However, the client says it cannot find /var/remoteboot/tftpboot/pxelinux.0. Removing the tftp-root= setting in dnsmasq.conf makes the client get the correct /tftpboot/pxelinux.0 so the problem seems to be in dnsmasq. It's behaving as designed: You've set the TFTP root to be /var/remoteboot, so filenames are relative to that root. The client asks for /tftpboot/pxelinux.0 so dnsmasq tries to send /var/remoteboot/tftpboot/pxelinux.0 which doesn't exist, so it returns an error, which inlcudes a message giving the complete pathname. That's what the client it displaying. Note that if the filename includes a leading /, dnsmasq will also try assuming it's an absolute pathname, but only if the first part of the filename matches the tftp-root. In a first look I didn't find any places where this concatenation could happen. Note that the tftp server is not the machine running dnsmasq in this case. Now I'm confused. What is the TFTP server? That's the whole point I don't understand. Here are the options for the machine: ometepe,vl5,3 ometepe,vl5,6 ometepe,vl5,28 ometepe,vl5,15 ometepe,vl5,66,192.168.5.74 ometepe,vl5,67,/tftpboot/pxelinux.0 The IP of the machine running dnsmasq is 192.168.5.18. So what should happen is client broadcasts dhcp request 192.168.5.18 answers saying tftp server is 192.168.5.74 client asks /tftpboot/pxelinux.0 to 192.168.5.74 client never heards about /var/remoteboot... That's why I gave the log line with the value of boot-filename above. Hmm... Looking at the log again the whole transaction is: DHCP packet: transaction-id is 3866001293 Available DHCP subnet: 192.168.5.1/255.255.255.0 Vendor class: PXEClient:Arch:0:UNDI:002001 DHCPREQUEST(eth0.5) 192.168.5.71 00:1e:8c:7f:6e:e6 DHCPACK(eth0.5) 192.168.5.71 00:1e:8c:7f:6e:e6 ometepe requested options: 1:netmask, 2:time-offset, 3:router, 5, 6:dns-server, requested options: 11, 12:hostname, 13:boot-file-size, 15:domain-name, requested options: 16:swap-server, 17:root-path, 18:extension-path, requested options: 43:vendor-encap, 54:server-identifier, 60:vendor-class, requested options: 67:bootfile-name, 128, 129, 130, 131, 132, requested options: 133, 134, 135 server name: 192.168.5.74 tags: vl5, ometepe, known sent size: 1 option: 53:message-type 05 sent size: 4 option: 54:server-identifier c0:a8:05:12 sent size: 4 option: 51:lease-time ff:ff:ff:ff sent size: 4 option: 1:netmask ff:ff:ff:00 sent size: 7 option: 12:hostname 6f:6d:65:74:65:70:65 sent size: 21 option: 67:bootfile-name 2f:74:66:74:70:62:6f:6f:74:2f:70:78:65... So it seems the client isn't requesting option 66 and is asking the dhcp server for pxelinux. If this is the case would --dhcp-option-force help? Would the tftp server be available to pxelinux in the client? The dnsmasq log shows no TFTP request but I vaguely remember Simon saying that these are not logged. File-not-found is not logged, since it clutters up the log with lots of failed attempts by PXELinux to read possible config files. But it helps in situations like this one. One can always grep them out, and compression ratio is very high for repetitive parts so it won't fill the disk. I've come across many PXE ROMS which are buggy, and just don't work if the DHCP server and TFTP server are on different machines. My guess is that you have one of those. Simon.
[Dnsmasq-discuss] multiple domain support - question
I have support for multiple domains working, but I've come across a wrinkle. Consider the case that two different DHCP clients claim the same name. With the existing code, only one can have it and the current behaviour is that when a second machine claims a name, the first one loses it. Now, consider the possibility that the two machines claiming the same name are in different domains. By default, the existing behaviour must continue, because the unqualified name is added to the DNS, so that even though the two clients could have name.domain1.com and name.domain2.com, they are still fighting over just plain name. It would be possible to introduce a new mode, which didn't put the unqualified name into the DNS, and allowed both hosts to keep their name as long as they are in different domains. Would that be useful, or just an confusing complication? Cheers, Simon.
Re: [Dnsmasq-discuss] Problems with Samba shares and XP clients
On Aug 3, 2008, at 12:15 PM, Cristiano Paris wrote: On Sun, Aug 3, 2008 at 2:06 PM, B. Cook bc...@poughkeepsieschools.org wrote: On Aug 2, 2008, at 5:00 PM, Cristiano Paris wrote: On Sat, Aug 2, 2008 at 6:30 PM, richardvo...@gmail.com richardvo...@gmail.com wrote: I don't remember for sure, but I don't think you need netbios for cifs to work. Try \\ip.address\share ? If that doesn't work then I don't think the problem is with dnsmasq. It doesn't work, niether. If that does work, you may have to mess with dns settings on both client and server until ping servername works, then \\servername\share should also work. Of course, ping works. Thank you. Cristiano You wouldn't want to share you config would you? dnsmasq and/or samba.. ___ Dnsmasq-discuss mailing list Dnsmasq-discuss@lists.thekelleys.org.uk http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss dnsmasq.conf: resolv-file=/etc/dns.resolv.conf local=/internal.theshire.org/ no-hosts addn-hosts=/etc/dns.hosts domain=internal.theshire.org dhcp-range=10.7.0.3,10.7.0.5,255.255.255.0,12h dhcp-host=00:xx:xx:xx:xx:xx,xx.xx.xx.xx,tuckburrow,infinite dhcp-option=44,10.7.0.2 dhcp-option=45,10.7.0.2 dhcp-option=46,8 dhcp-option=47 smb.conf: [global] workgroup = THESHIRE server string = The Shire internal server security = user load printers = yes log file = /var/log/samba/%m.log max log size = 50 wins support = yes name resolve order = wins lmhosts host bcast unix extensions = no socket options = TCP_NODELAY SO_RCVBUF=524288 SO_SNDBUF=524288 [homes] comment = Home Directories browseable = no writable = yes [printers] comment = All Printers path = /var/spool/samba browseable = no guest ok = no writable = no printable = yes Cristiano For what its worth.. 1 B-node: Broadcast - no WINS 2 P-node: Peer - WINS only 4 M-node: Mixed - broadcast, then WINS 8 H-node: Hybrid - WINS, then broadcast I believe that to be a correct and complete list for dhcp-option=46 (46 netbios-nodetype) I use 2, because I use wins. You might set your log level up a bit in samba and look in your log.nmbd and see if the hosts are actually registering with wins, is your samba server at 10.7.0.2? This is for my samba server @home.. # Global parameters [global] workgroup = WORKGROUP netbios name = THINGONE server string = [SAMBA] log file = /var/log/samba/log.%m name resolve order = wins bcast host lmhosts time server = Yes socket options = TCP_NODELAY IPTOS_LOWDELAY IPTOS_THROUGHPUT SO_KEEPALIVE SO_RCVBUF=8192 SO_SNDBUF=8192 logon script = logon path = logon home = domain logons = Yes os level = 65 preferred master = Yes local master = Yes domain master = Yes dns proxy = No wins support = Yes read only = No writeable = Yes case sensitive = No veto oplock files = /*.doc/*.xls/*.mdb/ security = user load printers = yes dns proxy = no load printers = yes printing = cups printcap name = cups and some dnsmasq: bind-interfaces bogus-priv cache-size=1024 dhcp-authoritative dhcp-host=00:01:e6:4b:e3:2e,172.16.64.10,jetdirect dhcp-host=00:0d:93:ed:75:dc,,iancook dhcp-host=00:12:3f:7e:b8:10,172.16.64.19,cswebdevel dhcp-host=00:17:f2:c7:c4:4a,172.16.64.11,imac dhcp-host=00:19:e3:d3:fa:a9,,macbookpro dhcp-host=00:c0:f4:f5:e9:34,172.16.64.1,gw dhcp-leasefile=/var/db/dnsmasq.leases dhcp-option=option:dns-server,0.0.0.0 # DNS dhcp-option=option:domain-name,at.home # Domain dhcp-option=option:netbios-dd,0.0.0.0 # netbios datagram distribution server dhcp-option=option:netbios-nodetype,2 # netbios node type dhcp-option=option:netbios-ns,0.0.0.0 # set netbios-over-TCP/IP nameserver(s) aka WINS server(s) dhcp-option=option:netbios-scope # empty netbios scope. dhcp-option=option:ntp-server,172.16.64.1 # NTP dhcp-option=option:router,172.16.64.1 # Gateway dhcp-option=vendor:MSFT,2,1i dhcp-range=172.16.64.11,172.16.64.25,255.255.255.224,1h domain-needed domain=at.home except-interface=lo0 filterwin2k group=nobody interface=fxp0 log-dhcp log-facility=/var/log/dnsmasq.log no-negcache Again all works great.. (Running FreeBSD 7 on the dnsmasq and samba machines)
Re: [Dnsmasq-discuss] Missing something obvious #known
On Aug 3, 2008, at 4:32 PM, Simon Kelley wrote: B. Cook wrote: Hello All, I thought all of this was working as it should.. but apparently it is not.. from the dnsmasq.log: Aug 1 11:42:56 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored Aug 1 11:43:11 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored Aug 1 11:43:14 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored Aug 1 11:43:21 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored and from the config: 20 # 100 stuff 21 dhcp-option=net:phsalt,option:router,10.20.100.1 22 dhcp-range=phsalt,10.20.100.2,10.20.100.254,255.255.255.0,5m 23 dhcp-ignore=net:phsalt,#known 24 25 dhcp-host=00:0e:7f:3b:ba:2d,10.20.100.3,net:phsalt # someprinter 26 dhcp-host=00:30:6e:c6:13:e6,10.20.100.57,net:phsalt # hplj1150_hs143 I have tried dhcp-ignore without the net: as well and it does not seem to cause an error either way.. I am trying to only let certain hosts end up in the 100.0/24.. Not sure what I am missing.. can anyone help? You don't need to use dhcp-ignore to achieve what you're trying to do, the net:phsalt in the dhcp-range is enough to ensure that the range only gets used for the hosts which set the phsalt tag. However, I'm not sure _why_ the configuration you have is behaving as it does. Please could you set --log-dhcp and send me the information that dnsmasq logs during these DHCP transactions? Cheers, Simon. So you are saying I should be able to have a single dhcp-ignore=#known and that would take care of it for all the subnets? [/usr/local/etc/dnsmasq]# 25 grep dhcp-range * dnsmasq.cli.dhcp:dhcp-range=cli,10.20.18.2,10.20.19.254,255.255.254.0,4h dnsmasq.kri.dhcp:dhcp-range=kri,10.20.4.2,10.20.5.254,255.255.254.0,4h dnsmasq.phs.dhcp:dhcp-range=phs,10.20.0.17,10.20.1.254,255.255.254.0,4h dnsmasq.phs.dhcp:dhcp-range=phsalt, 10.20.100.2,10.20.100.254,255.255.255.0,5m dnsmasq.smi.dhcp:dhcp-range=smi,10.20.6.2,10.20.7.254,255.255.254.0,4h dnsmasq.war.dhcp:dhcp-range=war,10.20.14.2,10.20.15.254,255.255.254.0,4h [/usr/local/etc/dnsmasq]# 26 grep \#known * dnsmasq.cli.dhcp:dhcp-ignore=cli,#known dnsmasq.kri.dhcp:dhcp-ignore=kri,#known dnsmasq.phs.dhcp:dhcp-ignore=#known dnsmasq.phs.dhcp:dhcp-ignore=phs,#known dnsmasq.phs.dhcp:dhcp-ignore=phsalt,#known dnsmasq.smi.dhcp:dhcp-ignore=smi,#known dnsmasq.war.dhcp:dhcp-ignore=war,#known
Re: [Dnsmasq-discuss] Missing something obvious #known
On Sun, August 3, 2008 4:32 pm, Simon Kelley wrote: B. Cook wrote: Hello All, I thought all of this was working as it should.. but apparently it is not.. from the dnsmasq.log: Aug 1 11:42:56 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored Aug 1 11:43:11 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored Aug 1 11:43:14 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored Aug 1 11:43:21 dnsmasq[32635]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d ignored and from the config: 20 # 100 stuff 21 dhcp-option=net:phsalt,option:router,10.20.100.1 22 dhcp-range=phsalt,10.20.100.2,10.20.100.254,255.255.255.0,5m 23 dhcp-ignore=net:phsalt,#known 24 25 dhcp-host=00:0e:7f:3b:ba:2d,10.20.100.3,net:phsalt # someprinter 26 dhcp-host=00:30:6e:c6:13:e6,10.20.100.57,net:phsalt # hplj1150_hs143 I have tried dhcp-ignore without the net: as well and it does not seem to cause an error either way.. I am trying to only let certain hosts end up in the 100.0/24.. Not sure what I am missing.. can anyone help? You don't need to use dhcp-ignore to achieve what you're trying to do, the net:phsalt in the dhcp-range is enough to ensure that the range only gets used for the hosts which set the phsalt tag. However, I'm not sure _why_ the configuration you have is behaving as it does. Please could you set --log-dhcp and send me the information that dnsmasq logs during these DHCP transactions? Cheers, Simon. So a few strange things.. I removed the dhcp-range=net:phsalt,#known, and still I was not able to get an ip in the 10.20.100.0/24 range. I removed the dhcp-range=#known (for the whole scope of the lan) and I was able to get an ip, dhcp-log'in showed me it was being accepted into the phs range, so thinking that somehow net:phs and net:phsalt was not being matched fully I :%s/phsalt/altphs/g in the configs.. and it is still being matched inside net:phs. Aug 4 09:11:01 dnsmasq[22625]: DHCP packet: transaction-id is 4133013384 Aug 4 09:11:01 dnsmasq[22625]: Available DHCP range: 10.20.0.17 -- 10.20.1.254 Aug 4 09:11:01 dnsmasq[22625]: Vendor class: Hewlett-Packard JetDirect Aug 4 09:11:01 dnsmasq[22625]: DHCPDISCOVER(em0) 00:0e:7f:3b:ba:2d Aug 4 09:11:01 dnsmasq[22625]: DHCPOFFER(em0) 10.20.0.213 00:0e:7f:3b:ba:2d Aug 4 09:11:01 dnsmasq[22625]: requested options: 1:netmask, 3:router, 44:netbios-ns, 6:dns-server, Aug 4 09:11:01 dnsmasq[22625]: requested options: 7:log-server, 12:hostname, 15:domain-name, Aug 4 09:11:01 dnsmasq[22625]: requested options: 22:max-datagram-reassembly, 54:server-identifier, Aug 4 09:11:01 dnsmasq[22625]: requested options: 58:T1, 59:T2, 69:smtp-server, 18:extension-path, Aug 4 09:11:01 dnsmasq[22625]: requested options: 144 Aug 4 09:11:01 dnsmasq[22625]: tags: phs Aug 4 09:11:01 dnsmasq[22625]: sent size: 1 option: 53:message-type 02 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 54:server-identifier 0a:14:00:10 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 51:lease-time 00:00:38:40 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 58:T1 00:00:1c:20 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 59:T2 00:00:31:38 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 1:netmask ff:ff:fe:00 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 28:broadcast 0a:14:01:ff Aug 4 09:11:01 dnsmasq[22625]: sent size: 8 option: 15:domain-name 70:68:73:2e:70:63:73:64 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 44:netbios-ns 0a:14:00:e6 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 3:router 0a:14:00:01 Aug 4 09:11:01 dnsmasq[22625]: sent size: 4 option: 6:dns-server 0a:14:00:10 something strange I also noticed, which might also help is this: (this is a different host) Aug 4 09:10:52 dnsmasq[22625]: DHCP packet: transaction-id is 1899741050 Aug 4 09:10:52 dnsmasq[22625]: Available DHCP range: 10.20.0.17 -- 10.20.1.254 Aug 4 09:10:52 dnsmasq[22625]: Vendor class: Hewlett-Packard JetDirect Aug 4 09:10:52 dnsmasq[22625]: DHCPREQUEST(em0) 10.20.0.186 00:1b:78:19:c9:50 Aug 4 09:10:52 dnsmasq[22625]: DHCPACK(em0) 10.20.0.186 00:1b:78:19:c9:50 hp4250-lib Aug 4 09:10:52 dnsmasq[22625]: requested options: 1:netmask, 3:router, 44:netbios-ns, 6:dns-server, Aug 4 09:10:52 dnsmasq[22625]: requested options: 7:log-server, 12:hostname, 15:domain-name, Aug 4 09:10:52 dnsmasq[22625]: requested options: 22:max-datagram-reassembly, 54:server-identifier, Aug 4 09:10:52 dnsmasq[22625]: requested options: 58:T1, 59:T2, 69:smtp-server, 18:extension-path, Aug 4 09:10:52 dnsmasq[22625]: requested options: 144 Aug 4 09:10:52 dnsmasq[22625]: tags: phs, phs, known Aug 4 09:10:52 dnsmasq[22625]: sent size: 1 option: 53:message-type 05 Aug 4 09:10:52 dnsmasq[22625]: sent size: 4 option: 54:server-identifier 0a:14:00:10 Aug 4 09:10:52 dnsmasq[22625]: sent size: 4 option: 51:lease-time 00:00:01:2c Aug 4 09:10:52 dnsmasq[22625]: sent size: 4 option: 58:T1 00:00:00:8a Aug 4
[Dnsmasq-discuss] question about logs
Is there a way to put the 'dns' logs somewhere and the 'dhcp' logs somewhere else? If it is not possible at this time.. Having so many requests come in for different things.. it would be nice to seperate them. my .02
Re: [Dnsmasq-discuss] illegal domain in dhcp-config directive
On Tue, Jul 29, 2008 at 07:40:52AM -0400, Jason Dixon wrote: On Tue, Jul 29, 2008 at 07:48:32AM +0100, Simon Kelley wrote: Jason Dixon wrote: Thanks to the recent vulnerability, I'm upgrading an older dnsmasq system from version 2.34 to 2.45. I've installed it and attempted to run it, but I get the following error. dnsmasq: illegal domain 04-06.lab.foobar.com in dhcp-config directive. The same configuration works fine with 2.34, so I presume something has changed in the configuration syntax or default settings. I haven't found anything on Google or in the FAQ about this error, nor do I see any dhcp-config directives in our config files or the sample config. Here is the relevant line from our config: dhcp-host=00:0c:29:02:1f:06,ubuntu-8.04-06.lab.foobar.com,172.16.0.105 It doesn't appear to be anything specifically with that line. We have a couple hundred of these loaded by dnsmasq. If I comment out that one, the next preceding one will error. I'm not experienced with dnsmasq, so I'm putting myself at the mercy of this list. Any help will be greatly appreciated. :) As far as I can see, this error can only be occuring because the domain part of the name your applying to that host doesn't match the domain specified for dnsmasq to use by the domain=domain config option. AFAIK dnsmasq has never allowed you to send a FQDN to a host unless it matches the domain setting (this is to avoid hosts impersonating, eg www.microsoft.com) but the check in this case has been moved from run-time, where it will log an error, to a fatal error at start-up The fix may be as simple as adding domain=04-06.lab.foobar.com (and you can remove the domain parts from all the dhcp-host lines as well, if you wish. If they stay, they must match the domain given) We have mixed domains in our configuration. I see the following line in the main dnsmasq.conf, which would account for most of the dhcp-host entries: local=/lab.foobar.com/wifi.foobar.com/ I see no domain directives anywhere. I suspect that adding 04-06.lab.foobar.com to the local directive might be sufficient, but I'm just guessing. And yes, this configuration works fine with dnsmasq-2.34. I tried various changes to our configuration but couldn't find a working set. I'm including a subset of our configuration here so you can see what might be causing the issue. The only bits I've left out include redundant information (e.g. dhcp-host). Just to reiterate, this works under 2.34 but is broken under 2.45. # cat /etc/dnsmasq.conf | grep -v '^$' | grep -v '^\#' filterwin2k local=/lab.foobar.com/wifi.foobar.com/ interface=eth0 expand-hosts dhcp-option=3,172.16.0.1 dhcp-option=42,0.0.0.0 conf-file=/etc/dnsmasq.lab.conf conf-file=/etc/dnsmasq.wifi.conf # cat /etc/dnsmasq.lab.conf | grep -v '^$' | grep -v '^\#' dhcp-range=lab,172.16.0.10,172.16.0.81,12h dhcp-option=lab,15,lab.foobar.com dhcp-host=00:0e:0c:bc:17:70,lab01.lab.foobar.com,172.16.0.10 dhcp-host=00:0c:29:ff:96:a7,solaris.lab.foobar.com,172.16.0.6 dhcp-host=00:0c:29:ff:96:xx,windows.lab.foobar.com,172.16.0.7 dhcp-host=00:0c:29:db:8a:b3,centos.lab.foobar.com,172.16.0.5 dhcp-host=00:0c:29:3c:58:01,centos-4.4-01.lab.foobar.com,172.16.0.82 # cat /etc/dnsmasq.wifi.conf | grep -v '^$' | grep -v '^\#' dhcp-range=wifi,172.16.0.200,172.16.0.250,3d dhcp-option=wifi,15,wifi.foobar.com Thanks, -- Jason Dixon OmniTI Computer Consulting, Inc. jdi...@omniti.com 443.325.1357 x.241
Re: [Dnsmasq-discuss] illegal domain in dhcp-config directive
Jason Dixon wrote: On Tue, Jul 29, 2008 at 07:40:52AM -0400, Jason Dixon wrote: On Tue, Jul 29, 2008 at 07:48:32AM +0100, Simon Kelley wrote: Jason Dixon wrote: Thanks to the recent vulnerability, I'm upgrading an older dnsmasq system from version 2.34 to 2.45. I've installed it and attempted to run it, but I get the following error. dnsmasq: illegal domain 04-06.lab.foobar.com in dhcp-config directive. The same configuration works fine with 2.34, so I presume something has changed in the configuration syntax or default settings. I haven't found anything on Google or in the FAQ about this error, nor do I see any dhcp-config directives in our config files or the sample config. Here is the relevant line from our config: dhcp-host=00:0c:29:02:1f:06,ubuntu-8.04-06.lab.foobar.com,172.16.0.105 It doesn't appear to be anything specifically with that line. We have a couple hundred of these loaded by dnsmasq. If I comment out that one, the next preceding one will error. I'm not experienced with dnsmasq, so I'm putting myself at the mercy of this list. Any help will be greatly appreciated. :) As far as I can see, this error can only be occuring because the domain part of the name your applying to that host doesn't match the domain specified for dnsmasq to use by the domain=domain config option. AFAIK dnsmasq has never allowed you to send a FQDN to a host unless it matches the domain setting (this is to avoid hosts impersonating, eg www.microsoft.com) but the check in this case has been moved from run-time, where it will log an error, to a fatal error at start-up The fix may be as simple as adding domain=04-06.lab.foobar.com (and you can remove the domain parts from all the dhcp-host lines as well, if you wish. If they stay, they must match the domain given) We have mixed domains in our configuration. I see the following line in the main dnsmasq.conf, which would account for most of the dhcp-host entries: local=/lab.foobar.com/wifi.foobar.com/ I see no domain directives anywhere. I suspect that adding 04-06.lab.foobar.com to the local directive might be sufficient, but I'm just guessing. And yes, this configuration works fine with dnsmasq-2.34. I tried various changes to our configuration but couldn't find a working set. I'm including a subset of our configuration here so you can see what might be causing the issue. The only bits I've left out include redundant information (e.g. dhcp-host). Just to reiterate, this works under 2.34 but is broken under 2.45. # cat /etc/dnsmasq.conf | grep -v '^$' | grep -v '^\#' filterwin2k local=/lab.foobar.com/wifi.foobar.com/ interface=eth0 expand-hosts dhcp-option=3,172.16.0.1 dhcp-option=42,0.0.0.0 conf-file=/etc/dnsmasq.lab.conf conf-file=/etc/dnsmasq.wifi.conf # cat /etc/dnsmasq.lab.conf | grep -v '^$' | grep -v '^\#' dhcp-range=lab,172.16.0.10,172.16.0.81,12h dhcp-option=lab,15,lab.foobar.com dhcp-host=00:0e:0c:bc:17:70,lab01.lab.foobar.com,172.16.0.10 dhcp-host=00:0c:29:ff:96:a7,solaris.lab.foobar.com,172.16.0.6 dhcp-host=00:0c:29:ff:96:xx,windows.lab.foobar.com,172.16.0.7 dhcp-host=00:0c:29:db:8a:b3,centos.lab.foobar.com,172.16.0.5 dhcp-host=00:0c:29:3c:58:01,centos-4.4-01.lab.foobar.com,172.16.0.82 # cat /etc/dnsmasq.wifi.conf | grep -v '^$' | grep -v '^\#' dhcp-range=wifi,172.16.0.200,172.16.0.250,3d dhcp-option=wifi,15,wifi.foobar.com Thanks, Are the names in your dhcp-host entries _all_ name.lab.foobar.com? If so, the simply adding domain=lab.foobar.com will fix things. I imagine that under 2.34 you will have seen lots of complaints logged about this, 2.45 has turned soft errors into hard ones. If there are names in more than one domain, the problem in more complicated to fix: you may need to try the new multi-domain code. Cheers, Simon.
Re: [Dnsmasq-discuss] illegal domain in dhcp-config directive
On Mon, Aug 04, 2008 at 04:16:41PM +0100, Simon Kelley wrote: Jason Dixon wrote: I tried various changes to our configuration but couldn't find a working set. I'm including a subset of our configuration here so you can see what might be causing the issue. The only bits I've left out include redundant information (e.g. dhcp-host). Just to reiterate, this works under 2.34 but is broken under 2.45. # cat /etc/dnsmasq.conf | grep -v '^$' | grep -v '^\#' filterwin2k local=/lab.foobar.com/wifi.foobar.com/ interface=eth0 expand-hosts dhcp-option=3,172.16.0.1 dhcp-option=42,0.0.0.0 conf-file=/etc/dnsmasq.lab.conf conf-file=/etc/dnsmasq.wifi.conf # cat /etc/dnsmasq.lab.conf | grep -v '^$' | grep -v '^\#' dhcp-range=lab,172.16.0.10,172.16.0.81,12h dhcp-option=lab,15,lab.foobar.com dhcp-host=00:0e:0c:bc:17:70,lab01.lab.foobar.com,172.16.0.10 dhcp-host=00:0c:29:ff:96:a7,solaris.lab.foobar.com,172.16.0.6 dhcp-host=00:0c:29:ff:96:xx,windows.lab.foobar.com,172.16.0.7 dhcp-host=00:0c:29:db:8a:b3,centos.lab.foobar.com,172.16.0.5 dhcp-host=00:0c:29:3c:58:01,centos-4.4-01.lab.foobar.com,172.16.0.82 # cat /etc/dnsmasq.wifi.conf | grep -v '^$' | grep -v '^\#' dhcp-range=wifi,172.16.0.200,172.16.0.250,3d dhcp-option=wifi,15,wifi.foobar.com Are the names in your dhcp-host entries _all_ name.lab.foobar.com? If so, the simply adding domain=lab.foobar.com will fix things. I imagine that under 2.34 you will have seen lots of complaints logged about this, 2.45 has turned soft errors into hard ones. If there are names in more than one domain, the problem in more complicated to fix: you may need to try the new multi-domain code. Yes, adding the domain=lab.foobar.com and changing the last section of dhcp-host's to remove the unwanted extra subdomain .4-0index. fixes the error. Thanks for your help! -- Jason Dixon OmniTI Computer Consulting, Inc. jdi...@omniti.com 443.325.1357 x.241
Re: My Emailiases: [Dnsmasq-discuss] multiple domain support - question
L Forrister wrote: Simon Kelley (si...@thekelleys.org.uk) (Emailias: REPLY-MASKED) wrote: Original Sender: dnsmasq-discuss-boun...@lists.thekelleys.org.uk Given To: lists.thekelleys.org.uk: Dnsmasq-discuss Info Page http://www.emailias.com/mailclick/?id=243457 end emailias header I have support for multiple domains working, but I've come across a wrinkle. Consider the case that two different DHCP clients claim the same name. With the existing code, only one can have it and the current behaviour is that when a second machine claims a name, the first one loses it. Now, consider the possibility that the two machines claiming the same name are in different domains. By default, the existing behaviour must continue, because the unqualified name is added to the DNS, so that even though the two clients could have name.domain1.com and name.domain2.com, they are still fighting over just plain name. It would be possible to introduce a new mode, which didn't put the unqualified name into the DNS, and allowed both hosts to keep their name as long as they are in different domains. Would that be useful, or just an confusing complication? I don't see anything confusing about it. What good is multiple domain support if you're forced to maintain unique host names across all domains. I've never quite understood why dnsmasq can't allow the duplicate names in the first place. Why must it be so. Consider a client machine with two interfaces. Both dhcp. They'll both get leases. Both ip's will be active. But the dns will only return one (name and ip) depending on which was most recently renewed. But its not just the dns, because when you check the leases file there's only one lease listed. The reason is simply that there's no way to differentiate the plain unqualified versions of the name, without the domain part. To make the multiple names work would mean suppressing unqualified names. For most people, having them is more usefull that having multiple domains. Cheers, Simon.
Re: My Emailiases: Re: My Emailiases: [Dnsmasq-discuss] multiple domain support - question
Simon Kelley (si...@thekelleys.org.uk) (Emailias: REPLY-MASKED) wrote: Original Sender: si...@thekelleys.org.uk Given To: lists.thekelleys.org.uk: Dnsmasq-discuss Info Page http://www.emailias.com/mailclick/?id=243457 end emailias header L Forrister wrote: Simon Kelley (si...@thekelleys.org.uk) (Emailias: REPLY-MASKED) wrote: Original Sender: dnsmasq-discuss-boun...@lists.thekelleys.org.uk Given To: lists.thekelleys.org.uk: Dnsmasq-discuss Info Page http://www.emailias.com/mailclick/?id=243457 end emailias header I have support for multiple domains working, but I've come across a wrinkle. Consider the case that two different DHCP clients claim the same name. With the existing code, only one can have it and the current behaviour is that when a second machine claims a name, the first one loses it. Now, consider the possibility that the two machines claiming the same name are in different domains. By default, the existing behaviour must continue, because the unqualified name is added to the DNS, so that even though the two clients could have name.domain1.com and name.domain2.com, they are still fighting over just plain name. It would be possible to introduce a new mode, which didn't put the unqualified name into the DNS, and allowed both hosts to keep their name as long as they are in different domains. Would that be useful, or just an confusing complication? I don't see anything confusing about it. What good is multiple domain support if you're forced to maintain unique host names across all domains. I've never quite understood why dnsmasq can't allow the duplicate names in the first place. Why must it be so. Consider a client machine with two interfaces. Both dhcp. They'll both get leases. Both ip's will be active. But the dns will only return one (name and ip) depending on which was most recently renewed. But its not just the dns, because when you check the leases file there's only one lease listed. The reason is simply that there's no way to differentiate the plain unqualified versions of the name, without the domain part. To make the multiple names work would mean suppressing unqualified names. For most people, having them is more usefull that having multiple domains. Cheers, Simon. As far the leases file is concerned, doesn't the mac address (and/or the client id) provide that differentiation? I mean, having the same unqualified host name on two separate leases would not cause identical lines/records in the lease file. The dns code, seems not to have a problem with multihomed hosts in either /etc/hosts or /etc/dnsmasq.hosts (my dnsmasq addn-hosts file.) oso:~ lf$ dig +short -x 192.168.160.67 -x 192.168.160.68 -x 192.168.160.69 alkix.rebel.lan. alkix.rebel.lan. alkix.rebel.lan. oso:~ lf$ dig +short alkix.rebel.lan 192.168.160.67 192.168.160.68 192.168.160.69 ~~L.Forrister
