Re: [Leaf-user] DCD vs. netsnmpd ???
Jeff Newmiller wrote: [snip] AFAIK, ld -s and strip -s should have the same effect, but if you omit the -s when you link (or compile/link, as with gcc) then you can debug it before you strip it for shipping. An interesting read is http://www-106.ibm.com/developerworks/library/l-shobj/ That's a really great link Jeff. Thanks a lot for the help. Matthew ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
On Fri, 8 Feb 2002, Matt Schalit wrote: Michael D. Schleif wrote: Is there a difference between these, especially regarding libraries? ld -s strip -s I'm sort of ignorant of the terminology, but I thought that strip was for the executable ELF files, and that ld -s would strip the ouput of the ld command, which is a shared library. Dave? Anyone? Some things I think I understand (in no particular order) are... ld is the gnu linker. binary executables are merged object files. To be executed, all of the unresolved symbols (e.g. printf) in a merged binary object must be resolved. (The right address for printf has to be stuffed into the object code where placeholders were put during compilation because printf is not typically in the .c file being compiled... it is in a library.) shared libraries are merged object files with symbols that can be referenced externally. executables may be linked as they are loaded with shared libraries that the system has been configured to use (ldconfig). They are usually statically linked to eliminate most of the unresolved symbols and insert housekeeping startup code, leaving only references to shared library symbols unresolved until the executable is loaded into memory. One use for a linker is to merge object files into shared libraries, and resolve the internal references between them. Another use for a linker is to merge object files into executables and resolve internal references between them. The main difference between these two uses is that the executable has a designated entry point at which execution may begin. There are two major types of superflous information that can be stripped out of an object file: extern symbols and source debug information. The former information must never be removed from a shared library, but both types can be removed from an excutable. AFAIK, ld -s and strip -s should have the same effect, but if you omit the -s when you link (or compile/link, as with gcc) then you can debug it before you strip it for shipping. An interesting read is http://www-106.ibm.com/developerworks/library/l-shobj/ but I think a better explanation of how this works under the hood ought to exist somewhere. --- Jeff NewmillerThe . . Go Live... DCN:[EMAIL PROTECTED]Basics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...2k --- ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
OK, I'm rolling my own. Yes, I know about strip for executables and this: # ./configure --prefix=/usr --enable-shared Wow! The libraries are 300% larger! What is the secret to minimizing space of shared libraries? What do you think? Michael D. Schleif wrote: netsnmpd.lrp (4.2.1-1-CS) from DCD v1.0.2 appears to be broken. [1] Changes to /etc/snmp/snmpd.conf do *not* affect snmpd. Specifically, modifying syscontact and syslocation are *not* accessible via snmpget nor snmpwalk, c. [2] Such (example) changes can be effected if snmpd is started with args: -c /etc/snmp/snmpd.conf [3] If arg -c is used, then snmpwalk will not end gracefully; rather: # snmpwalk -v 1 localhost public system system.sysDescr.0 = Linux trout 2.2.19-3-LEAF-RAID #4 Sat Dec 1 17:27:59 CST 2001 i486 . . . system.sysORTable.sysOREntry.sysORUpTime.9 = Timeticks: (5) 0:00:00.05 Timeout: No Response from localhost [4] In fact, no matter how the default startup is used, any community string has access to snmpd -- public, private or *none*! [5] Several attributes allowed here: http://net-snmp.sourceforge.net/man/snmpd.conf.html like this: sysname STRING *fail* on startup (using -c): # /etc/snmp/snmpd.conf: line 101: Warning: Unknown token: sysname. What do you think? -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
OK, I'm rolling my own. Yes, I know about strip for executables and this: # ./configure --prefix=/usr --enable-shared Wow! The libraries are 300% larger! What is the secret to minimizing space of shared libraries? You can strip libraries too... Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
Charles Steinkuehler wrote: OK, I'm rolling my own. Yes, I know about strip for executables and this: # ./configure --prefix=/usr --enable-shared Wow! The libraries are 300% larger! What is the secret to minimizing space of shared libraries? You can strip libraries too... Aha! Much, much better . . . Live and learn ; -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
I never did test that package, sorry. I put it together on the assumption that I would eventually use it, but that never happened. I apologize for the problems, I can build a better package if you would like, and verify that this one works first. If you do, please make note of the changes I made to your original package...notes are in the change log for Dachstein: http://leaf.sourceforge.net/devel/cstein/files/diskimages/dachstein-CD/chang es.txt You may not want all of the changes, but there are at least some that you'll probably want to include. Also, IIRC, I *did* try using the netsnmpd and netsnmpu packages, and was able to obtain the expected results. I don't have a handy test system at the moment, but I hope to try this again in a day or two. Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
Charles Steinkuehler wrote: I never did test that package, sorry. I put it together on the assumption that I would eventually use it, but that never happened. I apologize for the problems, I can build a better package if you would like, and verify that this one works first. If you do, please make note of the changes I made to your original package...notes are in the change log for Dachstein: http://leaf.sourceforge.net/devel/cstein/files/diskimages/dachstein-CD/chang es.txt You may not want all of the changes, but there are at least some that you'll probably want to include. Also, IIRC, I *did* try using the netsnmpd and netsnmpu packages, and was able to obtain the expected results. I don't have a handy test system at the moment, but I hope to try this again in a day or two. [1] I have rolled-my-own and it works. [2] You said: Broke single package into two pacakges: netsnmpd - snmp daemon netsnmpu - snmp utilities As I noted in another post: [3] net-snmp is really three (3) different pieces: snmp -- apps tools to do everything snmp snmpd -- facility to accept queries send traps snmptrapd -- facility to receive snmp traps It seems to me that these ought to be kept in separate packages, albeit dependent on installing the snmpd package. Although, it is common to want to remotely manage a firewal/router, I find it doubtful that these ought to receive/manage incoming traps; nor, may it be too common to want to query other hosts from the firewall? I suggest: netsnmp.lrp netsnmpd.lrp netsnmptrapd.lrp [3] I also have a working snmpd.conf, including working View-Based Access Control Model (vacm) examples. In fact, my snmpd.conf does *not* even mention community: public ; [4] A little more testing and Mike Noyes finalizing my leaf/sourceforge access, then I can post v2.4.3. What do you think? -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
I suggest: netsnmp.lrp netsnmpd.lrp netsnmptrapd.lrp How about netsnmptrapd - netsnmpt (still an 8 character limit on ms-dos fomatted floppies : [3] I also have a working snmpd.conf, including working View-Based Access Control Model (vacm) examples. In fact, my snmpd.conf does *not* even mention community: public ; [4] A little more testing and Mike Noyes finalizing my leaf/sourceforge access, then I can post v2.4.3. Sounds good. Any idea what the problem was with the earlier package? Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
Charles Steinkuehler wrote: I never did test that package, sorry. I put it together on the assumption that I would eventually use it, but that never happened. I apologize for the problems, I can build a better package if you would like, and verify that this one works first. If you do, please make note of the changes I made to your original package...notes are in the change log for Dachstein: http://leaf.sourceforge.net/devel/cstein/files/diskimages/dachstein-CD/chang es.txt You may not want all of the changes, but there are at least some that you'll probably want to include. Also, IIRC, I *did* try using the netsnmpd and netsnmpu packages, and was able to obtain the expected results. I don't have a handy test system at the moment, but I hope to try this again in a day or two. Further review of my new net-snmp package: [1] Only the .SO files in /usr/lib are required? [2] Anybody know if these can be needed and where, on a LEAF system? mib2c.conf mib2c.storage.conf mib2c.vartypes.conf [3] Do *not* need the header files under /usr/include/ucd-snmp ? [4] How did you strip description text from the mibs? Do you have a sed script? What do you think? -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
Charles Steinkuehler wrote: I suggest: netsnmp.lrp netsnmpd.lrp netsnmptrapd.lrp How about netsnmptrapd - netsnmpt (still an 8 character limit on ms-dos fomatted floppies : OK [3] I also have a working snmpd.conf, including working View-Based Access Control Model (vacm) examples. In fact, my snmpd.conf does *not* even mention community: public ; [4] A little more testing and Mike Noyes finalizing my leaf/sourceforge access, then I can post v2.4.3. Sounds good. Any idea what the problem was with the earlier package? I gave up trying to move the snmpd.conf under /etc/snmp. I tried various modifications to configure, config.h and Makefile. Everytime I tested a new scenario, it would *not* acknowledge snmpd.conf configurations, unless I started snmpd with -c -- then, although I could walk the mib, I was _deluged_ Authentication_Failure traps! I know that this can be done; but, how valuable is that? Currently, I am working with $prefix=/usr (default: /usr/local), which places most everything under here: /usr/share/snmp What do you think? -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
Michael D. Schleif wrote: Charles Steinkuehler wrote: Michael D. Schleif wrote: What is the secret to minimizing space of shared libraries? You can strip libraries too... Aha! Much, much better . . . Live and learn ; I though you knew you could strip libraries but didn't know how. Look at the man page for ld: -s Strip symbolic information from the output file. The debug and line sections and their associated relocation entries will be removed. Except for relocatable files or shared objects, the symbol table and string table sections will also be removed from the output object file. As it may not be the version of ld you're using, you might check that. I don't do this enough to be a big help, but I figure you got it by now :) Matt ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
Is there a difference between these, especially regarding libraries? ld -s strip -s Matt Schalit wrote: Michael D. Schleif wrote: Charles Steinkuehler wrote: Michael D. Schleif wrote: What is the secret to minimizing space of shared libraries? You can strip libraries too... Aha! Much, much better . . . Live and learn ; I though you knew you could strip libraries but didn't know how. Look at the man page for ld: -s Strip symbolic information from the output file. The debug and line sections and their associated relocation entries will be removed. Except for relocatable files or shared objects, the symbol table and string table sections will also be removed from the output object file. As it may not be the version of ld you're using, you might check that. I don't do this enough to be a big help, but I figure you got it by now :) Matt -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] DCD vs. netsnmpd ???
netsnmpd.lrp (4.2.1-1-CS) from DCD v1.0.2 appears to be broken. It's possible...this was a last minute addition, and is still considered experimental. I intend to switch to netsnmpd (and make some shell-script reporting agents for ipchains accounting rules), but have yet to do so. many problem details snipped [1] Changes to /etc/snmp/snmpd.conf do *not* affect snmpd. [2] Such (example) changes can be effected if snmpd is started with args: -c /etc/snmp/snmpd.conf [3] If arg -c is used, then snmpwalk will not end gracefully; rather: [4] In fact, no matter how the default startup is used, any community string has access to snmpd -- public, private or *none*! [5] Several attributes allowed here: http://net-snmp.sourceforge.net/man/snmpd.conf.html *fail* on startup (using -c): What do you think? Gak! Sounds like lots of problems...I thought I'd at least tried the default configuration, but maybe not. I'll do some tests and report back (probably take a day or two)... Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user