Menu
▾
▴
devmon-support
|
From: Andrew R. <an...@ei...> - 2009-08-11 18:44:15
|
Hello All, I recently updated our Cisco 3750 switches to the newest firmware from Cisco and ended up breaking my devmon tests for if_col, if_dsc, if_err, and if_load. It seems that Cisco has added SNMP monitoring of its "StackSub" ports (which I assume to be the stack interconnect ports on the back of the unit), in doing so they failed to add any information in the other statistics in the other branches about these ports. This leaves me "Undefined" in the table for those ports and the tests end up "clear". Is there a way in devmon to have a default value defined if no data is returned? In this case having the column be a default 0 would be preferable to an "undefined". - Andrew |
|
From: Richard F. <gol...@gm...> - 2009-08-11 20:39:31
|
On Tue, Aug 11, 2009 at 11:13, Andrew Rankin<an...@ei...> wrote:
> Is there a way in devmon to have a default value defined if no data is
> returned? In this case having the column be a default 0 would be
> preferable to an "undefined".
I, too, would like this. For Dell Poweredge servers running Windows
Server 2003, I'm separating out the RAID controllers into two tables
(layout, info) because an added host adapter ("DELL AHA-39160
(AIC-7899A) Ultra160 SCSI Host Adapter") doesn't provide firmware or
cache info.
|
|
From: Buchan M. <bg...@st...> - 2009-08-12 07:38:03
|
On Tuesday, 11 August 2009 21:38:55 Richard Finegold wrote:
> On Tue, Aug 11, 2009 at 11:13, Andrew Rankin<an...@ei...> wrote:
> > Is there a way in devmon to have a default value defined if no data is
> > returned? In this case having the column be a default 0 would be
> > preferable to an "undefined".
>
> I, too, would like this. For Dell Poweredge servers running Windows
> Server 2003, I'm separating out the RAID controllers into two tables
> (layout, info) because an added host adapter ("DELL AHA-39160
> (AIC-7899A) Ultra160 SCSI Host Adapter") doesn't provide firmware or
> cache info.
I don't think this is the same problem. Are you ending up with just "undefined"
in some columns of your table, or is the whole test failing due to "missing
repeater" ?
We have a situation like this, for the HP ProLiant's, where if an array has no
spare drive, the whole branch falls off ... so there are two templates, compaq-
server and compaq-servernohspare . However, in this case, I would prefer
devmon to distinguish between the "no response" and "empty response" cases.
However, this may be more easily fixed along with many other issues, by
changing SNMP modules, which I am still investigating.
Regards,
Buchan
|
|
From: Buchan M. <bg...@st...> - 2009-08-12 07:33:12
|
On Tuesday, 11 August 2009 19:13:49 Andrew Rankin wrote: > Hello All, > > I recently updated our Cisco 3750 switches to the newest firmware from > Cisco and ended up breaking my devmon tests for if_col, if_dsc, > if_err, and if_load. It seems that Cisco has added SNMP monitoring > of its "StackSub" ports (which I assume to be the stack interconnect > ports on the back of the unit), in doing so they failed to add any > information in the other statistics in the other branches about these > ports. This leaves me "Undefined" in the table for those ports and > the tests end up "clear". Does it not make more sense to exclude these ports from the affected tests for this device? This is what was discussed on the Hobbit list recently. > Is there a way in devmon to have a default value defined if no data is > returned? In this case having the column be a default 0 would be > preferable to an "undefined". I don't think devmon should be guessing when it doesn't know, however if there is no other solution, we can look at what options are available ... Regards, Buchan |
|
From: Andrew R. <an...@ei...> - 2009-08-13 20:31:14
|
Sorry for the slow reply. If someone wouldn't mind showing me how to exclude specific ports, I'm OK with that. What was the title of the thread on the hobbit list? For the time being I modified dm_tests.pm to not go 'clear' if it hits an 'Undefined', this allows hobbit to keep updating the remaining rrds. Andrew On Wed, Aug 12, 2009 at 3:32 AM, Buchan Milne<bg...@st...> wrote: > On Tuesday, 11 August 2009 19:13:49 Andrew Rankin wrote: >> Hello All, >> >> I recently updated our Cisco 3750 switches to the newest firmware from >> Cisco and ended up breaking my devmon tests for if_col, if_dsc, >> if_err, and if_load. It seems that Cisco has added SNMP monitoring >> of its "StackSub" ports (which I assume to be the stack interconnect >> ports on the back of the unit), in doing so they failed to add any >> information in the other statistics in the other branches about these >> ports. This leaves me "Undefined" in the table for those ports and >> the tests end up "clear". > > Does it not make more sense to exclude these ports from the affected tests for > this device? This is what was discussed on the Hobbit list recently. > >> Is there a way in devmon to have a default value defined if no data is >> returned? In this case having the column be a default 0 would be >> preferable to an "undefined". > > I don't think devmon should be guessing when it doesn't know, however if there > is no other solution, we can look at what options are available ... > > Regards, > Buchan > |
|
From: Andrew R. <an...@ei...> - 2009-08-13 20:37:02
|
Oop, I found it: Edit the following: /templates/cisco-3750/if_*/exceptions ifName : ignore : Nu.+|Vl.+|St.+ Andrew On Thu, Aug 13, 2009 at 4:30 PM, Andrew Rankin<an...@ei...> wrote: > Sorry for the slow reply. > > If someone wouldn't mind showing me how to exclude specific ports, I'm > OK with that. What was the title of the thread on the hobbit list? > > For the time being I modified dm_tests.pm to not go 'clear' if it hits > an 'Undefined', this allows hobbit to keep updating the remaining > rrds. > > Andrew > > On Wed, Aug 12, 2009 at 3:32 AM, Buchan Milne<bg...@st...> wrote: >> On Tuesday, 11 August 2009 19:13:49 Andrew Rankin wrote: >>> Hello All, >>> >>> I recently updated our Cisco 3750 switches to the newest firmware from >>> Cisco and ended up breaking my devmon tests for if_col, if_dsc, >>> if_err, and if_load. It seems that Cisco has added SNMP monitoring >>> of its "StackSub" ports (which I assume to be the stack interconnect >>> ports on the back of the unit), in doing so they failed to add any >>> information in the other statistics in the other branches about these >>> ports. This leaves me "Undefined" in the table for those ports and >>> the tests end up "clear". >> >> Does it not make more sense to exclude these ports from the affected tests for >> this device? This is what was discussed on the Hobbit list recently. >> >>> Is there a way in devmon to have a default value defined if no data is >>> returned? In this case having the column be a default 0 would be >>> preferable to an "undefined". >> >> I don't think devmon should be guessing when it doesn't know, however if there >> is no other solution, we can look at what options are available ... >> >> Regards, >> Buchan >> > |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X