Bug: WAN (PPPoE) Uptime incorrectly reset

[Saran]

I've noticed that with changing some settings (e.g. QoS Download), the "WAN (PPPoE) Uptime" value on the status page gets reset even though the PPPoE connection was not reset.

Uptime appears reset:

Uptime incorrectly reset to 0

2015-01-14_233928.png (19.83 KiB) Viewed 8179 times

But the IP is the same:

Google

2015-01-14_234850.png (8.16 KiB) Viewed 8179 times

as was 16 hours ago when router was restarted:

Time when DynDNS was updated

2015-01-14_233904.png (4.06 KiB) Viewed 8179 times

(mind the PST-CET TZ difference)

What gives?


Setup:
FW: gargoyle_1.6.2-ar71xx-generic-tl-wr1043nd-v1-squashfs-factory
HW: TP-Link TL-WR1043N/ND v1

[ispyisail]

when you make the change it would be good to check the "logread"

[Saran]

when you make the change it would be good to check the "logread"

So, here it is.

1. Status before:

status before

1.png (12.5 KiB) Viewed 8120 times

2. Change QoS upload and Save.

3. Status after:

status after

2.png (13.99 KiB) Viewed 8120 times

And here's the logread output:

Code: Select all

Jan 26 19:08:43 Gargoyle daemon.info dnsmasq-dhcp[6353]: DHCPINFORM(br-lan) <IP> <MAC> 
Jan 26 19:08:43 Gargoyle daemon.info dnsmasq-dhcp[6353]: DHCPACK(br-lan) <IP> <MAC> <HOST>
Jan 26 19:09:18 Gargoyle kern.info kernel: [1077585.310000] IMQ driver unloaded successfully.
Jan 26 19:09:19 Gargoyle kern.info kernel: [1077585.400000]     Hooked IMQ after mangle on INPUT
Jan 26 19:09:19 Gargoyle kern.info kernel: [1077585.410000]     Hooked IMQ after mangle on FORWARD
Jan 26 19:09:19 Gargoyle kern.info kernel: [1077585.410000] IMQ driver loaded successfully. (numdevs = 1, numqueues = 1)

I hope it helps...

[Saran]

Bump.

[ViBE]

This still happens. Some settings cause hangouts while applying.

[d3fz]

I'm experiencing the same issue, "WAN (PPPoE) Uptime" resets for no apparent reason, whenever i apply changes in random tasks (QoS Download/Upload, Add new Quota, Enable/Disable Ping Watchdog, etc).

Is this a bug or expected behavior ?

It doesn't hamper any functionality, just annoying when trying to monitor the uptimes.

[Lantis]

It's not really a bug.

When you make changes to fundamental parts of the firewall (e.g. qos, quotas, restrictions... anything really) the firewall needs to be restarted for these to take effect.

When the firewall restarts the ppp daemon also restarts, hence the reset time stamp.

[Saran]

Even if ppp deamon is restarted as you say, the connection is not reset, so the time shown is invalid --> bug.

Deamon needs to store the Uptime and restore it after restart is the connection is not changed. Otherwise it's displaying false information --> bug.

[Lantis]

Are you certain that the ppp daemon going down doesn't reset the connection?
I don't know that it does, but it seems logical to me. I don't have a pppoe connection anymore so I can't check the syslogs to confirm.

If you can think of a better way of determining up time of the PPPoE connection let me know and I'll look into it.
The use of the date for the pppd PID file is used widely from my research.

[Saran]

Check the WAN IP address from my before-after screenshots.