[prev in list] [next in list] [prev in thread] [next in thread] 

List:       net-snmp-patches
Subject:    [Net-snmp-patches] [ net-snmp-Patches-1102156 ] 5.1.2 jumbo patch
From:       "SourceForge.net" <noreply () sourceforge ! net>
Date:       2006-07-28 9:47:13
Message-ID: E1G6Ow5-0007fH-S8 () sc8-sf-web2 ! sourceforge ! net
[Download RAW message or body]

Patches item #1102156, was opened at 2005-01-14 08:03
Message generated for change (Comment added) made by dts12
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312694&aid=1102156&group_id=12694

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
Resolution: None
Priority: 5
Submitted By: Michael Kukat (vaxpower)
Assigned to: Nobody/Anonymous (nobody)
Summary: 5.1.2 jumbo patch for improved AIX 5.2 support

Initial Comment:
The attached patch enables a lot of functionality for 
AIX in net-snmp 5.1.2:

in mibgroup host, hrStorage is complete for all AIX 
versions now. hrSWRun is also useable, but information 
about path and parameters of processes don't come 
out, as i found no way to gather them.

MAC address in interfaces list blanked out, as i currently 
found no way to get the correct one. Counters are filled 
with the correct values now.

ip/icmp/tcp/udp statistics work on AIX 5.2 for the most 
common values, but still lots of values are missing. 
Especially the number of current established connection 
is missing, as libperfstat only provides the number of 
established connections since boot. On AIX 4.3.3 and 
AIX 5.1 all values are still missing, as libperfstat API 
changed to 5.2 and perfstat_protocol is not present in 
the previous versions. tcpstats and so are present in 
running kernel, but all values are zeroed.

The attached patch is currently not verified for 
influence with other platforms, this should be done.

Thanks go to Thomas Stark for defining the need and 
DaimlerChrysler for letting me work out those patches 
and allowing the contribution of the results to the 
community.

Results are welcome!

----------------------------------------------------------------------

>Comment By: Dave Shield (dts12)
Date: 2006-07-28 10:47

Message:
Logged In: YES 
user_id=88893

I've just applied Oliver's patch #1486307 to the main CVS
development code.  I'm assuming that this effectively
replaces the patches listed here?

Feel free to re-open this tracker entry if my assumptions
are wrong!

----------------------------------------------------------------------

Comment By: Michael Kukat (vaxpower)
Date: 2006-05-11 09:57

Message:
Logged In: YES 
user_id=794829

Thanks for the work, Oliver. Good to see that my work is 
kept alive in net-snmp, as i don't have the official 
chance to maintain it like it should be...

Michael

----------------------------------------------------------------------

Comment By: Oliver Lehmann (olivleh1)
Date: 2006-05-11 09:39

Message:
Logged In: YES 
user_id=344876

Find a patch for 5.3.1.pre1 (context diff) at

http://sourceforge.net/tracker/index.php?
func=detail&aid=1486307&group_id=12694&atid=312694

----------------------------------------------------------------------

Comment By: Thomas Anders (tanders)
Date: 2006-03-11 14:16

Message:
Logged In: YES 
user_id=848638

while a plain diff may be useful for your AIX users, it's
not helpful to net-snmp developers in order to review your
patch. Please attach a unified/context diff also.

----------------------------------------------------------------------

Comment By: Michael Kukat (vaxpower)
Date: 2005-12-02 13:04

Message:
Logged In: YES 
user_id=794829

After having a quick look at the bug database, i decided 
to attach out current patch for net-snmp 5.1.2, which 
provides a quite robust AIX support.

This again is plain diff format usable with AIX patch 
command, as this shall mainly help the users to get some 
working net-snmp on AIX. The difference to the last diff 
are just some 64bit issues, which i already saw in newer 
versions (5.2.1 or so) done for amd64 port. There might 
always be several endianess problems, which don't appear 
on amd64, i will dive into this next year with a more 
current version.

In short: If you want net-snmp on AIX, use 5.1.2 with 
these patches, if you encounter any problems with other 
versions.
And due to libperfstat, net-snmp must be compiled in 64bit 
on 64bit machines, otherwise the agent will hang or crash 
on get/walk on some performance related OIDs.

----------------------------------------------------------------------

Comment By: Michael Kukat (vaxpower)
Date: 2005-12-02 12:49

Message:
Logged In: YES 
user_id=794829

Took a while for me to have the time to look into this 
again. I'm still the maintainer of our project around net-
snmp, where AIX is still an important platform. We dropped 
AIX 4.3.3 from out internally supported platforms, 5.1 
also is currently not maintained, but the rest is. There 
were some other minor issues for 64bit i found, but as the 
base 5.1.2 we use is not really current, i would re-check 
everything with a more current version next year, when a 
package upgrade is scheduled (where i also want to upgrade 
net-snmp).
I think this would also be a good point to have a look at 
the current AIX bugs, as i start with the base net-snmp 
package than, without patching it first. This is more 
useful to check the bugs, as i can then insert my comments 
reflecting some unpatches sf.net version of the source.

----------------------------------------------------------------------

Comment By: Thomas Anders (tanders)
Date: 2005-08-19 09:35

Message:
Logged In: YES 
user_id=848638

The most effective way to help would be:
- use a fresh CVS checkout (MAIN or V5_2_patches)
- (re)submit your patch against this version

Also, it'd be extremely helpful to get your opinion on the
open AIX bugs in our bug DB. We don't have access to AIX, so
it's hard for us to deal with them. Search for Status==Open
and Group==aix and add your comments to those bugs,
especially the nlist problem (that I think is already
fixed). Thanks in advance!

----------------------------------------------------------------------

Comment By: Michael Kukat (vaxpower)
Date: 2005-08-19 05:29

Message:
Logged In: YES 
user_id=794829

Besides the machines at work (running 5.2 and 5.3 today,
mixed 32 and 64 bit), i have enough machines at home to test
4.3.3 and 5.1, usually also 5.2 is not a problem, as soon as
i get my 7043-150 back from a friend.

BTW, i recently found a 64bit issue. The AMD64 patches in
newer versions also have to be used for AIX (otherwise
strange Counter32 values :), and 32bit builds can't be run
on 64bit machines (libperfstat issue, can't be fixed, one
needs to build 2 versions if he needs 32 and 64 bit support).

Want to check all this in a newer version, so i'll send a
patch for some 5.2.x if necessary.

----------------------------------------------------------------------

Comment By: Thomas Anders (tanders)
Date: 2005-08-19 00:29

Message:
Logged In: YES 
user_id=848638

Do you still have access to AIX just in case we'll ask for a
test?

----------------------------------------------------------------------

Comment By: Michael Kukat (vaxpower)
Date: 2005-05-13 10:23

Message:
Logged In: YES 
user_id=794829

Here comes a unified diff. This is not applicable on default 
AIX installations, as patch doesn't know the format. GNU diff 
required. Hope, this helps the developers here.

----------------------------------------------------------------------

Comment By: Wes Hardaker (hardaker)
Date: 2005-05-06 06:38

Message:
Logged In: YES 
user_id=76242

Can you resubit your patch using diff -u instead of a straight diff?  
It's very hard to apply a diff created the way you submitted it to a 
moving target (ie, CVS trees).  diff -u gives a diff output that is 
easier to both read and apply. 

----------------------------------------------------------------------

Comment By: Michael Kukat (vaxpower)
Date: 2005-02-04 11:51

Message:
Logged In: YES 
user_id=794829

Replaced the patch due to a memory leak in hrStorage. This 
version is intensively tested, no further leaks detected on 
lots of repeated walks oder .1


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312694&aid=1102156&group_id=12694

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Net-snmp-patches mailing list
Net-snmp-patches@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/net-snmp-patches
[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic