AC86U Memory Usage (81992)

[Smedley]

Is it normal for the memory usage to drift upwards over time?

I just upgraded my AC86U running as an Access Point to version 81992. Memory usage after the upgrade was at 53%, and after 1 day and 18hours it's up to 56%.

I've read a number of threads here regarding how memory is used as a cache, but even overnight with little/no wireless traffic, memory usage increased. I can't say whether the same behavior happened on prior firmware versions.

Is this normal, or might it be indicative of a memory leak?

Thanks!

 

[ForkWNY]

That's fairly normal and is what I always observed when I used the AC86U as primary router. After a reboot the memory usage was lowest (reboot clears the memory), over time it would increase. I had the AC86U on a reboot schedule once a week which would clear out the memory every 7 days.

Over time I would think that memory caching/buffering based on the processes, network activity, client connections, and apps running within the AC86U's firmware would cause the usage to increase. I see the same behavior on my GT-AC5300 as well.

Any computer system you're using, especially Windows, you'll probably notice an uptick in memory usage the longer you run your system between reboots. As processes and apps use the RAM to cache data, memory usage will increase. Memory leaks typically cause erratic increases in RAM usage and may lend to instability. Programmers typically identify and fix memory leaks, patches/updates are released when they're fixed (though sometimes new ones are introduced). Always a good idea to keep your firmware updated as the benefits usually outweigh the drawbacks.

 

[OzarkEdge]

Is it normal for the memory usage to drift upwards over time?

I just upgraded my AC86U running as an Access Point to version 81992. Memory usage after the upgrade was at 53%, and after 1 day and 18hours it's up to 56%.

I've read a number of threads here regarding how memory is used as a cache, but even overnight with little/no wireless traffic, memory usage increased. I can't say whether the same behavior happened on prior firmware versions.

Is this normal, or might it be indicative of a memory leak?

Thanks!

From my lay observations...

I'd say your numbers are not yet indicative of a 'memory leak'. A memory leak will likely appear much more 'out of control.'

OE

 

[ColinTaylor]

Memory will be used for file/disk buffers, not network traffic. As you're using it as an AP presumably it doesn't have a USB drive plugged into it. So there will be a very small amount of buffering of the internal storage for things like log files, client lists, firmware updates. So unless the memory usage is going up very fast I don't think there's anything to worry about.

 

[tallytr]

For me its usually between 71% after reboot to 82% after a couple weeks....

 

[128bit]

i run a 1 tb ssd nas out of the 3.0 slot. the meter pretty much stays at 93%. i personally like that it's close to max from a utilization point of view. there are times we hit the nas fairly hard for tunes and vids. so far so good with interruptions being rare. i have read where the 2.0 slot is recommended; not sure why, so i've stayed this course.

 

[jsbeddow]

i have read where the 2.0 slot is recommended; not sure why, so i've stayed this course.

Easy to answer: using the USB 3.0 connection usually results in more wireless interference (radio noise). I know that on my 86U, using the USB 2.0 port I get a consistent "noise floor" level of -82 dBm to -90 dBm on the Wireless Log page. If I switch the same external UGreen external enclosure (for a small SSD) over to the USB 3.0 port, I will see noise readings consistently "higher" on the order of -10 dBm (ie, approximately -72 dBm to -78 dBm), so more background noise for the WiFi signal to deal with. Better shielded cables are said to reduce this effect, but I haven't experimented with that.

 

[128bit]

Easy to answer: using the USB 3.0 connection usually results in more wireless interference (radio noise). I know that on my 86U, using the USB 2.0 port I get a consistent "noise floor" level of -82 dBm to -90 dBm on the Wireless Log page. If I switch the same external UGreen external enclosure (for a small SSD) over to the USB 3.0 port, I will see noise readings consistently "higher" on the order of -10 dBm (ie, approximately -72 dBm to -78 dBm), so more background noise for the WiFi signal to deal with. Better shielded cables are said to reduce this effect, but I haven't experimented with that.

your tech response is much appreciated. i had read as much in some articles that were several years old and figured shielding would've have been improved by now. that said, it would be nice to "see" how the noise affects one's connection i.e., packet loss and/or signal degradation. my situation seems to be ok and a simple ping may be able to provide enough info for my home use.

 

[Smedley]

On 81992 running as an access point my memory usage is increasing at about 1MB every 5 hours.

Memory will be used for file/disk buffers, not network traffic. As you're using it as an AP presumably it doesn't have a USB drive plugged into it. So there will be a very small amount of buffering of the internal storage for things like log files, client lists, firmware updates. So unless the memory usage is going up very fast I don't think there's anything to worry about.

Memory usage is increasing by 1MB about every 5 hours while running in Access Point mode. The client list is fairly static, the log file is quiet, and there have been no firmware updates. I think there is a slow memory leak happening in 81992. Using the stock firmware, I don't see anyway to determine how/where the memory is being consumed. Are there commands I can issue through SSH that would provide more insight?

 

[ColinTaylor]

Using the stock firmware, I don't see anyway to determine how/where the memory is being consumed. Are there commands I can issue through SSH that would provide more insight?

Yes it's tricky to track down memory leaks, especially on stock firmware. I would start by looking at the output of:

1) free
2) df -h
3) top and then press M to sort by memory size.

But you'll have to wait until a significant proportion of the memory is being used to stand a chance of spotting something anomalous.

 

[128bit]

again, i'm running at a steady 93% for at least a few releases now. some may say that's high but i'm rolling along fine. 2 tvs streaming 4k, along with hd streams on the desktop. so, in a very dumb user sort of way, i'm in a "no worries" mode.
most of these guys here are much more advanced than me and i go way back to those days when networking was called tp (teleprocessing). i've had some basic packet training, token ring and very heavy tcp/ip training and have come to respect that some of this stuff is magic, i.e. proprietory.
compared to linksys, asus is a dream especially at these prices. all that said, these cheap soho routers could all use some tightening with security but i'm quite pleased with this overall setup. sounds like you're working to me. be happy

 

[Smedley]

Yes it's tricky to track down memory leaks, especially on stock firmware. I would start by looking at the output of:

1) free
2) df -h
3) top and then press M to sort by memory size.

But you'll have to wait until a significant proportion of the memory is being used to stand a chance of spotting something anomalous.

Many thanks! I'll check it periodically as the memory usage rises.