DOCSIS Signal Levels reporting "nan"

[rallen]

Hi,

I have a Motorola Surfboard 5101 that I'd like to keep an eye on. It seems that the version of Cacti that came in the installer had BSOD2600's DOCSIS stats files already built in - the ss_docsis.stats.php file was already in place and the appropriate data template was already available. I set up the graph just like I did for all the other equipment I have in there so far, and have the data sources mapped out correctly (I presume), but the graph doesn't appear to be taking any data. I have verified that the server Cacti resides on can reach that cable modem, and network usage graphs are registering properly. Is there anything I can check here to see what the problem is?

[BSOD2600]

snmpwalk .1.3.6.1.2.1.10.127.1. If you don't get any data, your ISP is likly blocking snmp requests from the LAN side.

[rallen]

I performed an snmpwalk of the cable modem and got tons of data back on it. Anything in particular you want me to look at, or are you just verifying that I can walk the modem?

[BSOD2600]

Primarly just to verify you could retrieve data. Do you see the basic snmp info in the top left corner in Cacti?

Look in the ss_docsis.stats.php file at the OIDs which it queries against. In your snmpwalk, you get all of those back with data, right?

Try manually running the script server and see if the script returns data.

[rallen]

Yes, when I look at the device properties I do see the SNMP Information:

Code: Select all

Lyons Office (10.11.12.141)
SNMP Information
System: HW_REV: 1; VENDOR: Motorola Corporation; BOOTR: 2164; SW_REV: SB5101-2.4.1.5-SCM01-NOSH; MODEL: SB5101
Uptime: 242223200
Hostname:
Location:
Contact:

All of the OIDs referenced in ss_docsis_stats.php are present in the results from the walk:

Code: Select all

transmission.127.1.1.1.1.6.3 = INTEGER: -120
transmission.127.1.1.4.1.5.3 = INTEGER: 346
transmission.127.1.1.4.1.6.3 = INTEGER: 33
transmission.127.1.2.2.1.4.2 = Counter32: 42
transmission.127.1.2.2.1.5.2 = Counter32: 1
transmission.127.1.2.2.1.10.2 = Counter32: 28
transmission.127.1.2.2.1.11.2 = Counter32: 0
transmission.127.1.2.2.1.12.2 = Counter32: 1944
transmission.127.1.2.2.1.13.2 = Counter32: 10

I looked at this modem in Stormwatch and it is reporting the same values (which I think would be expected, but still good to confirm anyway).

Here is what I noticed when I went to test the script server.

Code: Select all

Action: 2, Script: c:/Inetpub/wwwroot/cacti/scripts/ss_docsis_stats.php ss_docsis_stats 10.11.12.141public2161500

It looks like all of the parameters were run together instead of being spaced out. It didn't work as-is, but I spaced everything out and got the intended output (spaced for readability, it was all run together in the server):

Code: Select all

docsIfDownChannelPower:-121
docsIfSigQSignalNoise:348
docsIfSigQMicroreflections:33
docsIfCmRangingTimeout:5
docsIfCmStatusTxPower:420
docsIfCmStatusResets:42
docsIfCmStatusLostSyncs:1
docsIfCmStatusT1Timeouts:28
docsIfCmStatusT2Timeouts:0
docsIfCmStatusT3Timeouts:1944
docsIfCmStatusT4Timeouts:10

[BSOD2600]

What do you mean 'run together'? Please post exactly what was outputted from the script at the command line. The script should've put a single space between each group of text:values.

Also, change the logging level in cacti to debug for a cycle. Paste the relivant sections to this docsis script running.

[rallen]

What do you mean 'run together'? Please post exactly what was outputted from the script at the command line. The script should've put a single space between each group of text:values.

That's what it did, sorry for the confusion there.

I noticed something in the data input method for the DOCSIS stats which I corrected.

Code: Select all

<path_cacti>/scripts/ss_docsis_stats.php ss_docsis_stats <hostname><snmp_community><snmp_version><snmp_port> <snmp_timeout><snmpv3_auth_username><snmpv3_auth_password>

These were not spaced out properly, so I changed it to:

Code: Select all

<path_cacti>/scripts/ss_docsis_stats.php ss_docsis_stats <hostname> <snmp_community> <snmp_version> <snmp_port> <snmp_timeout> <snmpv3_auth_username> <snmpv3_auth_password>

Now I'm getting some data on the graph, but downstream still shows nan.

Attached is a snippet from the log run just now (after I made the above modification). It seems as though the program is getting the correct data, just having an issue with applying it to the graph?

[BSOD2600]

Hmm, interesting. I'll have to take a look at the template when I get home and make sure those variables are spaced out properly.

As for no downstream data, thats likely because in the data template, I've set the min/max too small. Increase it and then either delete the rrd file or use rrdtool tune to increase the existing datasource field.

[rallen]

Success! Changed the downstream range to -250 - 250 (I wouldn't have gone that far, except I've seen some customer modems at -200 somehow) and deleted the rrd file, now we're good to go. Wonder if this is another Windows 2000 burp

[rallen]

Hi, this issue has popped back up again.

From the cacti.log file in DEBUG mode:

Code: Select all

09/22/2006 11:25:55 AM - CMDPHP: Poller[0] Host[109] SNMP: Host responded to SNMP
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] DEBUG: SQL Exec: "update host set  status = '3',  status_event_count = '0',  status_fail_date = '2006-09-22 11:01:07',  status_rec_date = '0000-00-00 00:00:00',  status_last_error = 'Host did not respond to SNMP',  min_time = '9.99999',  max_time = '27.99000',  cur_time = '25.55',  avg_time = '19.0611111111',  total_polls = '10',  failed_polls = '1',  availability = '90'  where hostname = '10.10.1.10'"
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] DEBUG: SQL Assoc: "select     poller_reindex.data_query_id,     poller_reindex.action,     poller_reindex.op,     poller_reindex.assert_value,     poller_reindex.arg1     from poller_reindex     where poller_reindex.host_id=109"
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] Host[109] RECACHE: Processing 1 items in the auto reindex cache for '10.10.1.10'.
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] DEBUG: SQL Exec: "update poller_reindex set assert_value='199800' where host_id='109' and data_query_id='1' and arg1='.1.3.6.1.2.1.1.3.0'"
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] Host[109] DS[813] WARNING: Result from SERVER not valid.  Partial Result: 
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] Host[109] DS[813] SERVER: c:/Inetpub/wwwroot/cacti/scripts/ss_docsis_stats.php ss_docsis_stats 10.10.1.10 public 2 161 500  , output: U
09/22/2006 11:25:55 AM - CMDPHP: Poller[0] DEBUG: SQL Exec: "insert into poller_output (local_data_id,rrd_name,time,output) values (813,'','2006-09-22 11:25:00','U')"

I tested the script through the script server and it worked just fine. All other cable modems I have set up in Cacti are producing the same results. Is there any way to figure out what the server response actually is, instead of just "output: U"?

[BSOD2600]

Hi, this issue has popped back up again.

From the cacti.log file in DEBUG mode:

Code: Select all

09/22/2006 11:25:55 AM - CMDPHP: Poller[0] DEBUG: SQL Exec: "update host set  status = '3',  status_event_count = '0',  status_fail_date = '2006-09-22 11:01:07',  status_rec_date = '0000-00-00 00:00:00',  status_last_error = '[b]Host did not respond to SNMP[/b]',  min_time = '9.99999',  max_time = '27.99000',  cur_time = '25.55',  avg_time = '19.0611111111',  total_polls = '10',  failed_polls = '1',  availability = '90'  where hostname = '10.10.1.10'"

I tested the script through the script server and it worked just fine. All other cable modems I have set up in Cacti are producing the same results. Is there any way to figure out what the server response actually is, instead of just "output: U"?

1) You sure cacti properly see's the device as online? Looks like it was down before.
2) Is host 109 = 10.10.1.10?
3) when you ran the script with the script server, did all fields return values?
4) Try increasing the timeout to that device in cacti. Also wouldn't hurt to increase the max execution time and memory in php.ini (don't forget to restart IIS).

[rallen]

1) You sure cacti properly see's the device as online? Looks like it was down before.
2) Is host 109 = 10.10.1.10?
3) when you ran the script with the script server, did all fields return values?
4) Try increasing the timeout to that device in cacti. Also wouldn't hurt to increase the max execution time and memory in php.ini (don't forget to restart IIS).

1: Yes, all cable modems show as Up in the device list. Also, when I view an individual device I do see the SNMP information in the upper-left corner.

2: Yes.

3: Yes.

4: Memory and max execution time have already been bumped up quite a bit, I have memory set to 128m and execution time set to 3000. I will increase the SNMP timeout and let you know what happens.

Thanks!

[BSOD2600]

Hmm, sort of at a loss then why it wouldn't be returning any data. Try clearing the poller cache. You shouldn't need to adjust your php.ini file anymore, since your current settings are fine.

[rallen]

I increased the SNMP timeout and cleared the poller cache, then let it run over the weekend. No change in output.

[BSOD2600]

Alright, just to re-iterate things:

1) You can sucessfully snmpwalk 10.10.1.10 and get back the docsis data.
2) You can sucessfully run the docis script under the cacti script server and it returns data.
3) Having cacti run the script on 10.10.1.10 no data is returned.
4) Other cable modems of the exact same brand/type return docsis data with the script.