spdMerlin spdMerlin - Automated speedtests with graphs

[doczenith1]

The 19:40 test ran without issue.

Idle:

During the test when I saw the python entries populate:

I just ran a manual test. Grabbed a screenshot when I saw the memory shoot up. Then both putty windows froze and the system log shows "Apr 5 19:43:21 spdMerlin: Starting speedtest now..." with nothing after it. The test never finished.

Weird stuff happening. I went back to the router GUI and needed to log in even though it was showing the log page. After logging in the log no longer shows the "Apr 5 19:43:21 spdMerlin: Starting speedtest now..." entry. I now see:
Apr 5 19:40:00 spdMerlin: Starting speedtest now...
Apr 5 19:40:57 spdMerlin: Finished speedtest
Apr 5 19:40:57 spdMerlin: Speedtest results - Download: 570.76 Mbit/s - Upload: 10.70 Mbit/s - Ping: 28.62 ms
Apr 5 19:42:50 dropbear[4013]: Password auth succeeded for 'admin' from 192.168.1.4:61985
May 5 01:05:05 kernel: klogd started: BusyBox v1.25.1 (2019-04-03 18:32:13 EDT)

and then:
May 5 01:05:19 ntp: start NTP update
Apr 5 19:44:44 rc_service: ntp 1199:notify_rc restart_diskmon

I am pretty sure the router did not reboot but not certain. The uptime value does reflect this instance.

"Default message log level" is set to notice and "Log only messages more urgent than" is set to info

 

[Marin]

So how do I choose a server closest to me? I ran:

/jffs/scripts/spdcli.py --list

and I see all the servers I usually test with. How do I then select one of them?

@Jack Yaz, what command did you run to get the list below? It appears that it also gives you the option to select a server. Did I miss it on a previous post?

Also, is there a way to lock that server in the config file so it always used when running speed tests?

Thank You!

 

[Marin]

If it was my choice the top 5 would be great! Thank you!

I agree, 5 of the closest servers would be appropriate and plenty.

 

[doczenith1]

8:40 confirmed that the router rebooted. Nothing showing in log prior to reboot. I'm guessing that the previous time the router did reboot and I did not notice it.

 

[L&LD]

8:40 confirmed that the router rebooted. Nothing showing in log prior to reboot. I'm guessing that the previous time the routed did reboot and I did not notice it.

I have noticed that the RT-AC86U is very RAM hungry compared to all my previous Asus routers. Could it be because it is 64bit enabled?

With spdMerlin, our 1Gbps connections, and the woefully underpowered routers we have currently (any and all available, for consumers), I don't see this script being useful to us, as-is?

I would be curious if Jack Yaz could offer the 'single connection' testing option to see if it more closely matches an actual speedtest result (using the same parameters) from a more powerful device.

 

[QuikSilver]

Nope, not really. Just me and my laptop running speed tests from the spdmerlin when everyone is asleep.


Sent from my iPhone using Tapatalk

I'm noticing high pings as well on both routers (one at vacation home with no one there, while other shows consistent pings even during night). Low 59 high 89....Curious if location has to do with it.

 

[joe scian]

Works perfectly on AC-5300

 

[L&LD]

Works perfectly on AC-5300

What are, and does it match, your paid-for ISP speeds?

 

[JJohnson1988]

I have noticed that the RT-AC86U is very RAM hungry compared to all my previous Asus routers. Could it be because it is 64bit enabled?

I'm pretty sure the perceived high memory usage is nothing to worry about; it's just how the AC86U operates. I recall Merlin saying more data is cached in memory to speed things up.

 

[L&LD]

I'm pretty sure the perceived high memory usage is nothing to worry about; it's just how the AC86U operates. I recall Merlin saying more data is cached in memory to speed things up.

It's not perceived and I'm not worried either. But the underlying workings of the RT-AC86U are fundamentally different than any previous RMerlin powered router I've used. Just trying to understand it better.

And, it may be directly related to doczenith1's rebooting of his router too.

 

[joe scian]

What are, and does it match, your paid-for ISP speeds?

yes upload and download speed match - pings are a litte high 42ms compared to 26 ms when doing manual test at speedtest.net where the closest local server gives me 26ms ping. However its spot on when I choose dslreports.com/speedtest since I have 3 severs selected and 42ms is consistent with 42-44 ms reported by dslreports

 

[L&LD]

yes upload and download speed match - pings are a litte high 42ms compared to 26 ms when doing manual test

Thank you, and your ISP speeds?

 

[joe scian]

Thank you, and your ISP speeds?

I am on a 50D/20U FTTN connection. I get 45D and 15.4 U. Exactly as reported by Spdmerlin , Speedtest.net and DSLreports

 

[Jack Yaz]

I've updated post #1 with a link to https://github.com/sivel/speedtest-cli#inconsistency

Due to the nature of the device this is being run on, those with very high link speeds may see differing results to browser tests. There is nothing I can do about this. I will offer the --single switch in the next version to see if we can avoid outright crashes on high speed links, but please be aware you won't see your full link speed hit in the test.

 

[Jack Yaz]

So how do I choose a server closest to me? I ran:

/jffs/scripts/spdcli.py --list

and I see all the servers I usually test with. How do I then select one of them?

@Jack Yaz, what command did you run to get the list below? It appears that it also gives you the option to select a server. Did I miss it on a previous post?

Also, is there a way to lock that server in the config file so it always used when running speed tests?

Thank You!

It's a preview of my dev version. I spent several hours yesterday writing the code for selecting a server and being able to use your preferred server for the scheduled tests. I have more testing to do before its ready for release

 

[Jack Yaz]

8:40 confirmed that the router rebooted. Nothing showing in log prior to reboot. I'm guessing that the previous time the routed did reboot and I did not notice it.

Could you try something crazy like an 8GB swap file please? Or as large as you can go. Install htop from entware and search (with F3) for spdcli.py while the test is running. I'm curious to see how much memory is required. I expect your speeds will get hit by swapping coming in to play, however.

 

[Jack Yaz]

It's a preview of my dev version. I spent several hours yesterday writing the code for selecting a server and being able to use your preferred server for the scheduled tests. I have more testing to do before its ready for release

EDIT: This work is now going through internal testing (i.e. leave it running and listen for complaints of internet problems)

 

[Jack Yaz]

v1.1.0 is now available
Changelog:

Add the ability to choose a preferred server
Add the ability to choose between single and multi-connection mode

 

[Mutzli]

I noticed that since installing spdMerlin my free router memory is slowly declining. Yesterday evening it was at 312MB free and today around 1pm at 156MB. I just updated to 1.1.0. Let see how it looks after another 24h.

 

[L&LD]

I noticed that since installing spdMerlin my free router memory is slowly declining. Yesterday evening it was at 312MB free and today around 1pm at 156MB. I just updated to 1.1.0. Let see how it looks after another 24h.

You really should have a swap file installed. If you do, then free RAM is not an issue any longer.