Trouble with my gateway [solved but...]

[klui]

I am having problems with my gateway producing 404 errors after around 1000 queries to its web pages for my monitoring. I wanted to ask Sonic support if I could get another device to see if the problem is with the gateway or the firmware in general.

Support came back with the following:

Unfortunately we do not have a lot of options with regards to the modem and the Firmware is AT&T branded firmware so Pace doesn't really talk to us about it. I do encourage you to post to our member forums though to start discussion and see if any other customers can duplicate the issue or provide insight as to what resolution/mitigation there might be. I will check around a bit to see if I can duplicate the issue as well however our test line does not support that particular modem so I will have to find an alternative. I hope we can help you out. Thank you for choosing Sonic.

Since I'm paying Sonic for the gateway rental, I would presume I need to go through Sonic in order to arrange for a swap. Or do I go directly to AT&T? Confused.

EDIT: Subject edited
After updating the 5031nv with 10.5.1.525900-att, it looks like the 404 errors of the web server hasn't resurfaced. The "but..." in the edited subject refers to a new problem with this version of the firmware. The bytes transferred/received on pages C_1_0 and C_2_5 are now stored in a 32-bit variable, which means their values roll over to 0 after more than 4 GB of data are transferred.

I'm happy that the web server hasn't crashed, so far. Although it's only been a little less than 4.5 days since my test, I'm cautiously hopeful the server will stay running. The problem with the bytes transferred counters don't give me too much confidence; and, in fact, reinforces my prejudice that these firmware engineers don't care very much about what they do. When the next revision arrives, I am forced to worry what else would break.

[hhwong]

Why are you querying the gateway 1000 times?

[klui]

Why are you querying the gateway 1000 times?

I get DSL, session, transfer, and uptime stats from my modem every 5 minutes. After 3.5 days, the modem's web server takes a dump. I modify the interval to every 10 minutes and I get 404s after 7 days. I monitor most of my devices using SNMP at intervals between 5-10 minutes. AT&T's gateway is the only device I use web scraping. Even my old Cayman 3546 I could use telnet without any issues.

[hhwong]

I got the what you're doing - but I still don't know why. I guess AT&T has not been a trustworthy connection for you in the past?

If you are that savvy and sophisticated with your monitoring, why not just have your web scraper also reboot the RG in the middle of the night every couple of days?

[kgc]

FWIW, the NVG589 I have on my FTTN connection hasn't had any trouble. I've been scrapping the stats off the status page every 5 minutes since it was installed a couple of months ago.

[Guest]

I got the what you're doing - but I still don't know why. I guess AT&T has not been a trustworthy connection for you in the past? :-)

If you are that savvy and sophisticated with your monitoring, why not just have your web scraper also reboot the RG in the middle of the night every couple of days?

I started gathering them when my Cayman 3220 decided to lock up when its attenuation were below a certain level. It just happened all of a sudden after years of service. After the 3220 was upgraded to the 3546, I continued the monitoring as it also had the same behavior. I just wanted to continue after my FTTN migration. The http interface's nonce makes things a little more challenging but I think it is possible to do a programmatic reboot. My issue is I shouldn't need to do this. My 3546 locked up maybe once every couple of months and with the proper threshold functionality is restored automatically within 15 seconds. The 5031's braindead behavior just screams programming by bozos--or the result of management by bozos who wanted to remove SNMP/shell access.

KGC: what program do you use to monitor your 589's statistics? I spent a bit of time trying to determine whether if my problem was caused by the program but it's just a simple call and I don't see excessive ports being opened.

[dane]

I believe we had an issue too with our Pace gear, a memory leak that would eventually cause the onboard web server to fail. The unit would remain fully functional, just lose management via web. The team here worked with Pace and I believe this was subsequently resolved.

What's the make and specific model of the unit you have, and any firmware version data you can get from the interface?

[kgc]

KGC: what program do you use to monitor your 589's statistics? I spent a bit of time trying to determine whether if my problem was caused by the program but it's just a simple call and I don't see excessive ports being opened.

A very brute force perl script I threw together but I never got around to dumping it into anything useful.

https://microshaft.org/kgc/nvg589-kgc.pl

[klui]

I believe we had an issue too with our Pace gear, a memory leak that would eventually cause the onboard web server to fail. The unit would remain fully functional, just lose management via web. The team here worked with Pace and I believe this was subsequently resolved.

What's the make and specific model of the unit you have, and any firmware version data you can get from the interface?

Huh, now that's interesting. Maybe AT&T were never notified or they don't care...

Here's the a subset of the info that I included to support:

Code: Select all

Model					5031NV-030
Serial Number 		47141N144665
Hardware Version 	2701-000875-004
Software Version 	9.8.1.489233-att
Key Code				52HP-2374-2262-22AT-F2BQ
First Use Date		April 7, 2015
DSL Modem				A2pv6C037f4.d24c
System Password		Custom

Any suggestions on how to move forward would be greatly appreciated. Pace just punts support contact to AT&T if you're not a provider. Searches for the software version seems to indicate this is the most current version. I can't believe as a proper customer you guys are forced to use the web interface.

[klui]

KGC: what program do you use to monitor your 589's statistics? I spent a bit of time trying to determine whether if my problem was caused by the program but it's just a simple call and I don't see excessive ports being opened.

A very brute force perl script I threw together but I never got around to dumping it into anything useful.

https://microshaft.org/kgc/nvg589-kgc.pl

Thanks for that. The script I'm using references the same modules through Perl. BTW, if you're concerned about security, the gateway accepts SSL connections. You just need to download the certificate chain and reference it and LWP will do the right thing even if the device cert changes after each reboot. My guess is the device's lethargic boot time is due to it recreating the cert indicated by its serial number.

From my init.d script:

Code: Select all

HTTPS_CA_DIR=/root/lib/certs /root/bin/modem_stats.pl

Code: Select all

root@Ubuntu:~/lib/certs# ll
total 12
drwxr-xr-x 2 root root 4096 Apr 21 22:37 ./
drwxr-xr-x 3 root root 4096 Apr 21 22:37 ../
-rw-r--r-- 1 root root 2980 Apr 21 22:37 gateway.pace.com.pem
root@Ubuntu:~/lib/certs#