Toshiba e-Studio Copiers

[acacord]

Has anyone been able to get any useful/graphable data from a Toshiba e-Studio series copier? I'm mainly looking for page counts and toner levels.

-aa

[gandalf]

If you have the OIDs, I have the HowTos: http://docs.cacti.net/node/70
Reinhard

[acacord]

OK, here's what I have found so far....

I have 4 different models of Toshiba e-Studio copiers: 650, 4511, 451c, & 351c. The first is a black & white unit, while the latter 3 are color. The color units all use the same controller which is different from the B&W's.

Regarding SNMP on the machines, they all offer the ability to change the community. But, on the color units, if you change the community from the defaults ('public' & 'private'), the queries will fail. Basically, you break SNMP on those units if you modify the communities. The 650's are unaffected by this 'bug'.

The templates I'm using are:
- Printers - HP 2500 (host)
- Printer - HP - Black Toner Level (data)
- SNMP - HP Total Count (data)
- SNMP - Today Print (data)
- Printer - HP - Toner Level (graph)
- SNMP - Today Print Counter (graph)
- SNMP - Total Print Counter (graph)
All templates were downloaded from posts in this forum.

The 650's are reporting a value of 0 for the page count OID (43.10.2.1.4.1.1). Strange. I've talked to several people at a Toshiba subsidiary and they're going to help out with some internal legwork on this one. The same thing is happening with the toner level. Not a big deal, but would be nice.

The color units are a different story, as they are outputting the total impressions from the machine. Color, black, print, copy - it's all there. That's great, but not perfect. Unfortunately, I can't find a way to break out the color and black/white prints from the color units. Even though the devices track everything down to scans and counts by paper size, you can't get that information from SNMP. I think we may try to put together a perl script that curls the counters pages from the web interfaces of each machine and logs the values that we want for RRDTool to graph. I'm going to need some help from the community on this one, though.

The Today Print Counter is strange with these units. I think it resets the counter if I reset the controller. That, then hoses your counts. It seems pretty worthless to me, so I'll probably scratch those graphs.

That pretty much sums it all up, so far. I'll post more when I find it...

[gandalf]

The vendor of course should be able to supply the wanted OIDs (what else should snmp support of those device be for? I don't suppose they would support configuration by means of SNMP ...)
If not, the curl one is of course a possible workaround. Basics to work with them should be available from http://docs.cacti.net/node/70
Reinhard

[acacord]

Actually, they do use snmp for client configuration on Windows. That's why I broke the snmp queries when I changed the community on the color units. They apparently never designed in the ability to support other community strings.

Another fun fact: the 650/black and white machines use what Toshiba refers to as the GL-1020 controller. That device runs a Linux 2.4 kernel-based OS. That's all I can find out on it, though. I'll poke around more later... But on the color units, they use the "eBridge" controller, which used to run VxWorks...

[toffuit]

Purpose: to get page counter on those Toshiba e-Studio. tested on many model and they all use the same OIDs.

What I use:
- WinXP Pro
- xampplite
- Perl (full) from xampp package
- Net::Snmp for Perl <--- http://search.cpan.org/~dtown/Net-SNMP- ... et/SNMP.pm

Instruction: Follow the documentation for data input method.

[toffuit]

I use LAST CF. Make sure you set your CF properly in Graph template.

Here's the script and the graph output.

[BSOD2600]

While a Perl script doing snmpgets works.... it's quite inefficient. You really should either use stand-alone data templates with single OIDs, or create an SNMP XML script. Either would be loads faster.

[toffuit]

While a Perl script doing snmpgets works.... it's quite inefficient. You really should either use stand-alone data templates with single OIDs, or create an SNMP XML script. Either would be loads faster.

I appreciate the comment. I tried the snmp xml script, it worked but I didn't like the end result. It's messy and created too many duplicates poller cache entry. I think this is because the Toshiba do not have true index OID for these page counter OIDs.

But in saying that, there is a major weakness in netsnmp perl. I tried snmpget oid of different tree (or whatever it's called). Basically, it wont get oid when I do this:

Code: Select all

my $xxx= .1.3.6......1.1
my $yyy= .1.3.6......2.1
my $zzz = .1.3.6......3.1

it only works if

Code: Select all

my $xxx= .1.3.6......1.1
my $yyy= .1.3.6......1.2
my $yyy= .1.3.6......1.3

[BSOD2600]

One can use a REGEX in the snmp xml file to make/fake the Index for cacti. I've used the following trick for many recent templates/scripts: http://docs.cacti.net/howto:data_query_ ... _not_exist

Anyways, each to their own. Small scale, any scripting solution works. Just becomes an issue when scaling to more devices/queries