--> wes has joined #net-snmp
--> dts has joined #net-snmp
--> miike has joined #net-snmp
--> TrogL has joined #net-snmp
[2003/01/16 08:21:04] #net-snmp <TrogL> morning dave
[2003/01/16 08:21:29] #net-snmp <TrogL> getting back to our discussion, I've just discovered a bug in What's Up Gold, what I'm using for a MIB online browser
[2003/01/16 08:21:36] #net-snmp <TrogL> it can't handle non-contiguous requests
[2003/01/16 08:21:44] #net-snmp <TrogL> so I can't use it to graph my CPU usage
[2003/01/16 08:21:58] #net-snmp <TrogL> they say they're gonna fix it....but when?
[2003/01/16 08:23:09] #net-snmp <TrogL> this improves the argument to have our usage contiguous eg. RAM = 0, swap = 1, disk= 2+
[2003/01/16 08:23:22] #net-snmp <TrogL> even though linux delivers 1-3
[2003/01/16 08:25:06] #net-snmp <TrogL> incidentally, what is "Memory Buffers"? Why isn't it implemented on Solaris?
[2003/01/16 08:37:10] #net-snmp <dts> Sorry - I was away sorting out a problem.
[2003/01/16 08:37:31] #net-snmp <TrogL> np
[2003/01/16 08:37:40] #net-snmp <dts> Umm.... non-contiguous requests are perfectly valid, of course.
[2003/01/16 08:37:50] #net-snmp <TrogL> yeah I know, they admit it's broken
[2003/01/16 08:38:24] #net-snmp <dts> I had a quick look at the indexing of hrStorage - and it's currently closely linked to the hrDevice index
[2003/01/16 08:39:02] #net-snmp <dts> But it doesn't have to be - it'll just mean an extra lookup when implementing the linking object
[2003/01/16 08:39:07] #net-snmp <TrogL> incidentally, wes and others ... http://www.net-snmp.com/postcards/postcard.54.html this is the guy at WUG I'm working with
[2003/01/16 08:39:47] #net-snmp <dts> I've always been assuming that the index registries would be built upon 'snmp_enum'
[2003/01/16 08:40:03] #net-snmp <dts> But that doesn't seem to have any mechanism for walking through the enumerations.
[2003/01/16 08:40:16] #net-snmp <TrogL> hang on
[2003/01/16 08:40:22] #net-snmp <TrogL> ACTION prints source
[2003/01/16 08:40:35] #net-snmp <dts> wes: and/or Robert - what do you feel about:
[2003/01/16 08:40:59] #net-snmp <dts> a) Providing a 'first/next' style API for the snmp_enum stuff, and
[2003/01/16 08:41:09] #net-snmp <dts> b) ordering enumerations internally
[2003/01/16 08:41:36] #net-snmp <dts> Is this OK, would it break anything, or am I barking up the wrong tree completely?
[2003/01/16 08:42:45] #net-snmp <TrogL> I don't want to increase the overhead if possible
[2003/01/16 08:42:56] #net-snmp <TrogL> WUG is handling the indexes correctly
[2003/01/16 08:43:16] #net-snmp <TrogL> I think
[2003/01/16 08:43:39] #net-snmp <TrogL> ACTION looks
[2003/01/16 08:45:18] #net-snmp <dts> If you're looking for a short-term fix, then the easiest thing is probably to switch the order
[2003/01/16 08:45:29] #net-snmp <dts> of the "fixed" entries, and the dynamic ones.
[2003/01/16 08:45:43] #net-snmp <TrogL> nope, I'm wrong - it can't find the indexs
[2003/01/16 08:45:49] #net-snmp <dts> I.e. start with memory/swap/buffers, and then work through the disks.
[2003/01/16 08:45:56] #net-snmp <TrogL> yes, taht's what I intend
[2003/01/16 08:45:59] #net-snmp <dts> That would at least keep things contiguous.
[2003/01/16 08:46:40] #net-snmp <TrogL> I'd love to hard-code the disks starting at 4 just to make life easy, but I'm confused as to what linux is delivering for 103 "Memory Buffers"
[2003/01/16 08:47:33] #net-snmp <TrogL> but then again, WUG sounds like they want to fix it
[2003/01/16 08:47:53] #net-snmp <TrogL> ACTION fetches his printout
[2003/01/16 09:08:32] #net-snmp <dts> TrogL: I've got some good news, and some bad news.
[2003/01/16 09:08:46] #net-snmp <dts> The good news is that I've thrown together a patch to switch the order.
[2003/01/16 09:09:00] #net-snmp <dts> See http://www.csc.liv.ac.uk/~daves/hrstorage.diff
[2003/01/16 09:09:39] #net-snmp <dts> The bad news is that it stops reporting after hrStorageAllocationUnits
[2003/01/16 09:10:03] #net-snmp <dts> in other words, it skips size, used and allocation failures :-(
[2003/01/16 09:10:29] #net-snmp <dts> But it's a start.....
[2003/01/16 09:28:42] #net-snmp <dts> Time for me to wander homewards - but I'll have a look at fixing things properly tomorrow.
[2003/01/16 09:28:47] #net-snmp <dts> G'night all
--> sharuk has joined #net-snmp
--> jlixfeld has joined #net-snmp
[2003/01/16 11:19:39] #net-snmp <jlixfeld> Anyone got net-snmp-5.0.6 working with perl5.8.0? I get "Can't use string ("SNMP") as a HASH ref while "strict refs" in use at /usr/lib/perl5/5.8.0/ExtUtils/MM_Unix.pm line 541." when trying to compile...
[2003/01/16 11:19:50] #net-snmp <jlixfeld> er..
[2003/01/16 11:20:04] #net-snmp <jlixfeld> the snmp module, rather.. shit.. think I'm in the wrong place...
[2003/01/16 11:20:59] #net-snmp <jlixfeld> .. or not.. didn't know SNMP module came with net-snmp..
[2003/01/16 11:33:50] #net-snmp <wes> The SNMP module does come with net-snmp.
[2003/01/16 11:34:09] #net-snmp <wes> Try getting 5.0.7 and running ./configure --with-perl-modules in the net-snmp directory. (then make and make install)
[2003/01/16 11:40:19] #net-snmp <jlixfeld> will try.. thx.
[2003/01/16 11:40:28] #net-snmp <wes> is this on a sun?
[2003/01/16 11:40:31] #net-snmp <wes> with Sun's perl?
[2003/01/16 11:53:36] #net-snmp <jlixfeld> no.. pc.. rh8
--> TrogL has joined #net-snmp
[2003/01/16 11:56:03] #net-snmp <TrogL> stupid network
--> benr has joined #net-snmp
[2003/01/16 12:28:21] #net-snmp <TrogL> grrr.. on IRIX, I'm trying to use openssl
[2003/01/16 12:28:31] #net-snmp <TrogL> it's installed in a variety of places in /usr/freeware
[2003/01/16 12:28:42] #net-snmp <TrogL> how do I explain to configure how to find it
[2003/01/16 12:29:14] #net-snmp <TrogL> last attempt was --with-openssl="/usr/freeware/include/openssl"
[2003/01/16 12:29:23] #net-snmp <TrogL> becuase that's where i found the files it was looking for
[2003/01/16 12:29:36] #net-snmp <TrogL> configure: error: Asked to use OpenSSL but I couldn't find it
[2003/01/16 12:49:56] #net-snmp <wes> --with-openssl=/usr/freeware/openssl and the include and libs directories should be in there.
[2003/01/16 12:49:59] #net-snmp <wes> If it isn't that way:
[2003/01/16 12:50:35] #net-snmp <wes> --with-cflags=-I/usr/freeware/include --with-ldflags=-L/usr/freeware/lib
[2003/01/16 13:10:22] #net-snmp <TrogL> neither one worked. i'll look at it later
[2003/01/16 13:10:44] #net-snmp <TrogL> robert I've got builds of 2_6, 7 and 8 as per our last discussion
[2003/01/16 13:45:15] #net-snmp <TrogL> wes: I added -v like the guy in -codes said and I've got more detailed output, but it looks really sad
[2003/01/16 13:46:44] #net-snmp <TrogL> I'll cut back on my options and see if it clear up any
[2003/01/16 13:49:19] #net-snmp <wes> -v for what?
[2003/01/16 13:49:28] #net-snmp <wes> (I haven't read my -coders in a few hours)
[2003/01/16 13:49:39] #net-snmp <TrogL> cflags
[2003/01/16 13:49:46] #net-snmp <TrogL> he wanted more verbose C barf code
[2003/01/16 13:50:07] #net-snmp <wes> TrogL: mail me the binaries when you give them to robert so I can create a signature for them.
[2003/01/16 13:50:21] #net-snmp <TrogL> I don't know if my mailer can handle 56mb binaries
[2003/01/16 13:51:01] #net-snmp <wes> Ok, ignore it. I'll work directly with Robert.
[2003/01/16 13:51:32] #net-snmp <TrogL> is there an ftp site i can upload to? wanna upload from my ftp site?
[2003/01/16 13:51:43] #net-snmp <wes> that would wokr.
[2003/01/16 13:51:46] #net-snmp <wes> (the later)
[2003/01/16 13:51:49] #net-snmp <TrogL> lemme see if I've got room
[2003/01/16 13:52:09] #net-snmp <TrogL> I might have to do them one at a time
[2003/01/16 13:52:42] #net-snmp <wes> that'd be fine too
[2003/01/16 13:53:09] #net-snmp <TrogL> just gzipping them now
[2003/01/16 13:56:21] #net-snmp <TrogL> ACTION watches steam coming out of 3 development boxen
[2003/01/16 14:00:18] #net-snmp <TrogL> just running a quick check on them on a test box
[2003/01/16 14:05:00] #net-snmp <TrogL> nope, something wrong
[2003/01/16 14:08:47] #net-snmp <TrogL> brb
[2003/01/16 14:19:03] #net-snmp <TrogL> should I strip the executables? it might cut the size down a tad
[2003/01/16 14:22:37] #net-snmp <TrogL> hmmm /usr/local/sbin/snmpd syntax error at line 1: ( unexpected
[2003/01/16 14:22:53] #net-snmp <TrogL> they were built on an Ultra450 running solaris 2_7
[2003/01/16 14:23:11] #net-snmp <TrogL> I used the make install prefix=/foo etc. to build in a separate place
[2003/01/16 14:23:17] #net-snmp <TrogL> I tarred that up and moved it over
[2003/01/16 14:23:23] #net-snmp <TrogL> to /usr2/dist
[2003/01/16 14:23:37] #net-snmp <TrogL> then I cd'd to / and typed
[2003/01/16 14:23:49] #net-snmp <TrogL> tar -xvf /usr2/dis/whatever it is
[2003/01/16 14:24:03] #net-snmp <TrogL> then I started snmpd and it went blooey
[2003/01/16 14:24:04] #net-snmp <TrogL> any ideas?
[2003/01/16 14:37:36] #net-snmp <wes> its a script? (check)
[2003/01/16 15:00:52] #net-snmp <TrogL> nope - file reports...
[2003/01/16 15:01:30] #net-snmp <TrogL> ELF 32-bit MSB executable SPARC32PLUS Version 1, V8+ Required, UltraSPRAC1 Extensions required, dynamically linked, not stripped
[2003/01/16 15:01:49] #net-snmp <TrogL> it is NOT an Ultrasparc by any stretch of the imagination
[2003/01/16 15:02:56] #net-snmp <TrogL> these are different than the ones I compiled a few weeks ago
[2003/01/16 15:03:18] #net-snmp <TrogL> they are ELF 32-bit MSB executable SPARC Version 1, dynamically linked, not stripped
[2003/01/16 15:04:23] #net-snmp <wes> got me.
[2003/01/16 15:04:26] #net-snmp <wes> I'm not a sun expert.
[2003/01/16 15:04:30] #net-snmp <wes> (any more)
[2003/01/16 15:05:24] #net-snmp <TrogL> lemme try the 2_6 version
[2003/01/16 15:09:41] #net-snmp <TrogL> boom
[2003/01/16 15:10:06] #net-snmp <TrogL> ls.os.1 /usr/local/sbin/snmpd: fata: libc.so.1 version SUNW_1.18 not found
[2003/01/16 15:10:11] #net-snmp <TrogL> (required by file /ustr/local/bin/snmpd)_
[2003/01/16 15:10:30] #net-snmp <TrogL> guess we'll have to wait for robert
[2003/01/16 15:25:47] #net-snmp <TrogL> what's up gold people got back to me with a patch