- WLUG - wlug.mailman3.com

Re: [Wlug] RH7.3 Questions

by Charles R. Anderson

On Mon, Sep 02, 2002 at 10:48:55PM -0500, Richard Goodman wrote: dick> On Q2, yes I really mean to run dhcpd to issue IP#s to selected dick> client machines on the internal network. I'm confident that my dick> dhcpd.conf is correct, and it is called by "dhcpd eth1". This was dick> running under RH7.0. I installed the dhcpd RPM off the 7.3 CD (don't dick> have version info here) Ok, I just looked at the RH 7.3 dhcp package. You can put this in /etc/sysconfig/dhcpd: DHCPDARGS="eth1" Then set the service to start automatically with: /sbin/chkconfig dhcpd on and start it right now with: /sbin/service dhcpd start However, if you have specified eth1, it probably isn't starting due to problem #1. You would see a message from the DHCP server in /var/log/messages to that effect. dick> 1b) /sbin/ifconfig shows only "lo" and "eth0" and nothing for "eth1" Yes, but what does "/sbin/ifconfig eth1" show? By default, ifconfig without any arguments only shows interfaces that are UP. By specifying eth1 explicitly, it will print out that interface whether it is UP or DOWN. Alternatively, you can do "/sbin/ifconfig -a" to show all interfaces, UP or DOWN. dick> 1c) There is no /etc/modules.conf file on my machine. Strange. Is there a /etc/conf.modules? There really should be one of those two. Otherwise, eth0 most likely wouldn't work, either, unless you compiled it into the kernel. dick> 1d) I think the module is 8139too, because when I temporarily pulled dick> the Realtek card (eth1) I got a message during the attempt to startup dick> ETH1 complaining about that module by name (did not get complete dick> message). Ok, so /etc/modules.conf should have at least this in it: alias eth0 tulip alias eth1 8139too dick> I did some fiddling with modprobe (which I don't understand) after dick> reading section 2.3 of the Ethernet How-To (Using more than one dick> Ethernet Card per machine). From the startup post screen I can see dick> the Ethernet cards are on IRQs 10 & 9, and from the ifconfig I see dick> that eth0 is 0x8000, but I couldn't get the address for eth1. Do I dick> need it? These are PCI cards, right? Then you don't need any of that. The PCI BIOS autoconfigures the IRQ's and addresses. What you can do is: /sbin/lspci -v and check to see if your Realtek shows up there. In there it should also have IRQ's and addresses assigned. This command should get things going manually: /sbin/modprobe 8193too After that, check "dmesg" output and see if there are any error messages about loading 8139too. If not, "/sbin/ifup eth1" should start up the network on that interface. -- Charles R. Anderson <cra(a)wpi.edu> / http://angus.ind.wpi.edu/~cra/ PGP Key ID: 49BB5886 Fingerprint: EBA3 A106 7C93 FA07 8E15 3AC2 C367 A0F9 49BB 5886

Sept. 3, 2002

Server For Sale

by JV

Sometime around April I bought a server from Dell. The plan, at the time, was to make it into a webserver and host Wormtown.org and whoever else I could split the bandwidth/hard drive with at a hosting facility in Marlboro. However, it was also in April that I lost my job and then determined that the donations that Wormtown.org collects are not nearly enough to cover the cost of hosting. So, now that I've been unemployed since April. I'm feeling the money start to tighten up and I think it's about time to sell this server. It is a Dell PowerEdge 1650 http://www.dell.com/us/en/esg/topics/esg_pedge_rackmain_servers_1_pedge_165… It is a 1U rackmountable server. It has one PIII 1.13 chip in it It has 512M of RAM (2x256) It has one 36G drive (U160, 10K RPM) The rest of the specs are standard. I've got all the CDs including the Dell Drivers CD and the Red Hat 7.2 Discs that came with it. I've even got Rack Rails. And it does have the cool blue Dell logo that lights up on it. It's been run for about 15 hours total. I believe I paid a bit over $2100 for it. I just priced the same exact machine for $2071 (plus tax and shipping) at Dell.com. I will, of course, take the best offer I can get. I don't want to go any lower than $1500 if I can help it, though. If you're interested, or someone you know might be, contact me. Thanks John V +--------------------------------------------+ |-- "I taught you how to cyclops rock |-- (and then you go and) |-- turn around and break my heart |-- and waste my cyclops time |-- mess up my cyclops mind" |-- |-- - They Might Be Giants |-- |-- John: http://voegtlin.org/~john +--------------------------------------------+

Sept. 3, 2002

Re: [Wlug] RH7.3 Questions

by Richard Goodman

Charles, Thank you for your detailed response. Your answers to Q 3-7 close those issues. On Q2, yes I really mean to run dhcpd to issue IP#s to selected client machines on the internal network. I'm confident that my dhcpd.conf is correct, and it is called by "dhcpd eth1". This was running under RH7.0. I installed the dhcpd RPM off the 7.3 CD (don't have version info here) Since I first wrote, I went back to the new server and banged my head on it for a couple more hours, remarkably getting answers to some of your questions. 1a) ifcfg-eth1 does have "ONBOOT=yes". I will post it when I get back to the server in the morning. 1b) /sbin/ifconfig shows only "lo" and "eth0" and nothing for "eth1" 1c) There is no /etc/modules.conf file on my machine. 1d) I think the module is 8139too, because when I temporarily pulled the Realtek card (eth1) I got a message during the attempt to startup ETH1 complaining about that module by name (did not get complete message). I did some fiddling with modprobe (which I don't understand) after reading section 2.3 of the Ethernet How-To (Using more than one Ethernet Card per machine). From the startup post screen I can see the Ethernet cards are on IRQs 10 & 9, and from the ifconfig I see that eth0 is 0x8000, but I couldn't get the address for eth1. Do I need it? Thank you for your help; obviously Q1 is the key to getting running now, Dick >On Mon, Sep 02, 2002 at 06:45:59PM -0500, Richard Goodman wrote: >dick> 1) During startup eth0 comes up but eth1 does not. My notes on the >dick> previous installation [year ago?] don't show that I did anything >dick> special when I installed the Realtek (was previously also DEC 2104x) >dick> but that doesn't mean I just didn't write it down. > >Edit /etc/sysconfig/network-scripts/ifcfg-eth1 and make sure >"ONBOOT=yes" is in >there. Could you post that file so we can look at it? > >Run "/sbin/ifconfig eth1". What does that show? If it says "Device >not found" then >most likely the driver module isn't loading. If not, check >/etc/modules.conf. It >should have a line for eth1 like so: > >alias eth1 foobar > >where "foobar" is the name of the driver module for the Realtek. >There are a couple >different possibilites for the driver, since they made some different cards >with different chipsets. Try ne2k-pci and 8139too. > >You can also try running /usr/sbin/setup as root and going to "Network >configuration". > >dick> 2) dhcpd won't run, but I assume that's related to #1. Was surprised >dick> that dhcpd wasn't loaded during install, as I flagged eth1 as dhcp. > >What do you mean you flagged eth1 as dhcp? That sounds like you mean >"Use DHCP to get >the address for eth1". If that is the case, then you want a DHCP >*client*, not the >dhcpd server. That would be either pump or dhcpcd or dhclient. >pump, at the least, >should always be installed by Red Hat's installer. > >If you really wanted a DHCP server, then you would have to install >Red Hat's package >dhcp-2.0pl5-8.i386.rpm (or my packages for the latest ISC release >3.0.1rc9, available >at ftp://angus.ind.wpi.edu/pub/packages/isc/dhcp) You need to be >careful that you >configure the server correctly to only listen for requests on your >internal LAN eth1 >(for my packages, edit /etc/sysconfig/dhcpd and put INTERFACES=eth1 in there). > >-- >Charles R. Anderson <cra(a)wpi.edu> / http://angus.ind.wpi.edu/~cra/ >PGP Key ID: 49BB5886 >Fingerprint: EBA3 A106 7C93 FA07 8E15 3AC2 C367 A0F9 49BB 5886 > >_______________________________________________ >Wlug mailing list >Wlug(a)mail.wlug.org >http://mail.wlug.org/mailman/listinfo/wlug -- ------------------------------------------------- Richard Goodman dick(a)goodman1.net ---> Home: (508) 753-DICK

Sept. 3, 2002

Re: [Wlug] RH7.3 Questions

by Charles R. Anderson

On Mon, Sep 02, 2002 at 06:45:59PM -0500, Richard Goodman wrote: dick> 1) During startup eth0 comes up but eth1 does not. My notes on the dick> previous installation [year ago?] don't show that I did anything dick> special when I installed the Realtek (was previously also DEC 2104x) dick> but that doesn't mean I just didn't write it down. Edit /etc/sysconfig/network-scripts/ifcfg-eth1 and make sure "ONBOOT=yes" is in there. Could you post that file so we can look at it? Run "/sbin/ifconfig eth1". What does that show? If it says "Device not found" then most likely the driver module isn't loading. If not, check /etc/modules.conf. It should have a line for eth1 like so: alias eth1 foobar where "foobar" is the name of the driver module for the Realtek. There are a couple different possibilites for the driver, since they made some different cards with different chipsets. Try ne2k-pci and 8139too. You can also try running /usr/sbin/setup as root and going to "Network configuration". dick> 2) dhcpd won't run, but I assume that's related to #1. Was surprised dick> that dhcpd wasn't loaded during install, as I flagged eth1 as dhcp. What do you mean you flagged eth1 as dhcp? That sounds like you mean "Use DHCP to get the address for eth1". If that is the case, then you want a DHCP *client*, not the dhcpd server. That would be either pump or dhcpcd or dhclient. pump, at the least, should always be installed by Red Hat's installer. If you really wanted a DHCP server, then you would have to install Red Hat's package dhcp-2.0pl5-8.i386.rpm (or my packages for the latest ISC release 3.0.1rc9, available at ftp://angus.ind.wpi.edu/pub/packages/isc/dhcp) You need to be careful that you configure the server correctly to only listen for requests on your internal LAN eth1 (for my packages, edit /etc/sysconfig/dhcpd and put INTERFACES=eth1 in there). dick> 3) how do I know whether IPchains or IPtables is active, and how do I dick> switch. Considering the time involved so far, I think I'd rather dick> delay switching to IPTables and use my existing IPChains firewall for dick> now. Can I still hook it in at the end of rc.local? Red Hat defaults to IPchains, even now. You can put your ipchains statements into /etc/sysconfig/ipchains, and it should be executed automatically at boot. Note that the file format isn't a shell script, but rather the output of "/sbin/ipchains-save". You can run a shell script to set up the rules and then run "/sbin/service ipchains save" to save them for you in the right place. dick> 4) During install I did answer some firewalling questions (set to dick> medium, custom) -- where is the results of that dialog? /etc/sysconfig/ipchains Also, since the ipchains stuff is started as a system service, you need to make sure it is configured to start up automatically: /sbin/chkconfig ipchains on dick> 5) If I can use my existing IPChains firewall, that will take care of dick> IP Masquerade -- if not can I have some guidance as to what direction dick> to take. Well, does your existing IPChains firewall do IP Masquerade? If so, run the existing firewall script and do "/sbin/service ipchains save" to save it in /etc/sysconfig/ipchains. Note that this might not be flexible enough e.g. if you get a dynamic IP address from your ISP, in which case the rules might need to change slightly each time the firewall script is run. If you do use the above method, you'd need to make sure IP forwarding is also turned on in /etc/sysctl.conf. An alternative to using the /etc/sysconfig/ipchains and sysctl.conf stuff is to put your firewall script into /etc/sysconfig/network-scripts/ifup-post. After it runs the /sbin/ipchains commands, it would manually turn on IP forwarding with a "echo 1 > /proc/sys/net/ipv4/ip_forward". dick> 6) httpd comes up but says "cannot resolve hostname, using 127.0.0.1" dick> - what did I forget to configure? (I haven't touched httpd.conf yet.) Set your hostname in /etc/sysconfig/network, and put an entry for your host in /etc/hosts with the right IP address. dick> 7) issue and issue.net used to be set from rc.local, but no more -- dick> where are they set? I'd like to patch that out and be a bit more dick> anonymous. Yes, Red Hat has finally come to their senses and doesn't put anything in rc.local by default anymore. You can just edit /etc/issue and /etc/issue.net with a text editor, and they won't be touched by anything. -- Charles R. Anderson <cra(a)wpi.edu> / http://angus.ind.wpi.edu/~cra/ PGP Key ID: 49BB5886 Fingerprint: EBA3 A106 7C93 FA07 8E15 3AC2 C367 A0F9 49BB 5886

Sept. 3, 2002

Re: [Wlug] Importing email from Eudora or Outlook...

by Karl Hiramoto

Use an IMAP server. store your folders on the server. IMAP suports sub-folders. Then you access your e-mail independent of your mail client. On Sun, 1 Sep 2002, Stephen C. Daukas wrote: > I didn't see this covered in any of the posts I could read in the Mozill > groups, nor the FAQs, so here goes... > > I am trying to import email into Mozilla 1.0. I tried importing from > Eudora and got the account settings OK, but did not get the entire > address book (notably distribution lists) and couldn't get all the > email. I received no error for the address book, but did get an error > saying "cannot create proxy object for destination mailboxes" about half > way throuhg the import, at which point all the folder that were created > were moved to Trash. > > I decided to try Outlook just in case the Eudora code was messed up, I > first imported from Eudora into Outlook, but I got the same problem. > > In both cases, I get the impression that the import code can't do too > much with nested information (email distribution lists and mail folder > that contain other folders), but Mozilla seems to be able to get started > without any trouble. Then again, there is a lot of email involved and > I'm also wondering if there are some limits set somewhere in the code > before it gives up... > > Anyone have any suggestions? > > Thanks!! > > Steve > > > _______________________________________________ > Wlug mailing list > Wlug(a)mail.wlug.org > http://mail.wlug.org/mailman/listinfo/wlug > -- ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl(a)hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 Personal web page: http://karl.hiramoto.org/ Zoop Productions: http://www.zoop.org/ KTEQ Rapid City: http://www.kteq.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø The difference between dogs and cats is that dogs come when they're called. Cats take a message and get back to you.

Sept. 3, 2002

RH7.3 Questions

by Richard Goodman

I have a RH7.0 server which has been giving me regular unrecoverable disk errors since 6-8 weeks ago. I figured today I had enough time to do something about it. I also figured that as long as I had to rebuild from scratch (after installing a new hard drive) I'd upgrade to RH7.3 -- but I have a few more problems that I figured on. The server is dual homed - eth0 (DEC 2104x 10mbps tulip driver) goes to DSL, eth1 (Realtek 100mbps) goes to internal network. Was running IP Masquerade / IPChains / Apache / sendmail / DNS server fine under RH7.0 I have the following problems: 1) During startup eth0 comes up but eth1 does not. My notes on the previous installation [year ago?] don't show that I did anything special when I installed the Realtek (was previously also DEC 2104x) but that doesn't mean I just didn't write it down. 2) dhcpd won't run, but I assume that's related to #1. Was surprised that dhcpd wasn't loaded during install, as I flagged eth1 as dhcp. 3) how do I know whether IPchains or IPtables is active, and how do I switch. Considering the time involved so far, I think I'd rather delay switching to IPTables and use my existing IPChains firewall for now. Can I still hook it in at the end of rc.local? 4) During install I did answer some firewalling questions (set to medium, custom) -- where is the results of that dialog? 5) If I can use my existing IPChains firewall, that will take care of IP Masquerade -- if not can I have some guidance as to what direction to take. 6) httpd comes up but says "cannot resolve hostname, using 127.0.0.1" - what did I forget to configure? (I haven't touched httpd.conf yet.) 7) issue and issue.net used to be set from rc.local, but no more -- where are they set? I'd like to patch that out and be a bit more anonymous. I'm definitely unconnected until most of these Q's get resolved. Thanks to any who can assist me. Dick

Sept. 2, 2002

RE: [Wlug] WLUG's membership in BUG

by George Metz

I'm not a member, but I DID work at RCN during that MegaMeeting. Boston User Groups was started as a corporate-sponsored way to get US Datacenters into the User Group scene, as it were. Incidentally, US Datacenters is no longer in that building, IIRC. It sounds to me like the $50 could be better spent in a whole host of ways other than membership in a group for groups that don't interact much. Especially if you can't even get them to put events up on their calendar for your group. Of course, I MIGHT be biased by the fact that they took over the building I worked at for a meeting that they didn't clear with (or notify) any of the other companies OR the management in the building. At least the pizza was good. :)

Sept. 1, 2002

Importing email from Eudora or Outlook...

by Stephen C. Daukas

I didn't see this covered in any of the posts I could read in the Mozill groups, nor the FAQs, so here goes... I am trying to import email into Mozilla 1.0. I tried importing from Eudora and got the account settings OK, but did not get the entire address book (notably distribution lists) and couldn't get all the email. I received no error for the address book, but did get an error saying "cannot create proxy object for destination mailboxes" about half way throuhg the import, at which point all the folder that were created were moved to Trash. I decided to try Outlook just in case the Eudora code was messed up, I first imported from Eudora into Outlook, but I got the same problem. In both cases, I get the impression that the import code can't do too much with nested information (email distribution lists and mail folder that contain other folders), but Mozilla seems to be able to get started without any trouble. Then again, there is a lot of email involved and I'm also wondering if there are some limits set somewhere in the code before it gives up... Anyone have any suggestions? Thanks!! Steve

Sept. 1, 2002

Re: [Wlug] WLUG's membership in BUG

by Brian J. Conway

I vote no. We received no real benefits from our original 50 dollars spent, they won't list our events on their calendar, and their user group links don't even work through a text browser. That 50 bucks would be much better spent elsewhere (beer and pizza??). ;-) Perhaps people should just address their votes to Andy and not the list? -b

Aug. 30, 2002

Re: [Wlug] overlay/watermark???

by Keith Wright

> From: Bill Mills-Curran <subssn594(a)charter.net> > > I'm generating a document with good old latex, and I'd like to be able > to add a figure as the background to each page, something like a > watermark, or the background capability of html. > > I'd be happy to use a post-processing step to do this. I've been > looking at the ps utilities, but there's nothing that looks like it > will work. > > Any ideas? Make a PostScript graphic to fill the page, and use any graphic package you can find ( \usepackage{epsfig} works for me ) to tell TeX to put a tiny little box in the page header, which contains your big graphic. TeX doesn't care if your graphic (or anything else) is bigger than the box, it just needs the box to know how much space to leave, but you, you pervert, want to print something and leave no space for it. It should work fine in a technical sense, but promise me the background will be very light and sparse, or I shall revile and shun your document. -- Keith

Aug. 30, 2002