Create seperate log for Associations

[tkerns]

When I look at the log for V3 all I see is association and disassociations. All system info is lost. It would be great to break this into 2 different logs.... one system info and the other association info.

Tim Kerns
CV-Access

[tog]

You could send your logs to a syslog server so they're all available for review.

[tkerns]

and then flood syslog with several hundred association messages.... I prefer to have syslog contain real meaningful info.

[oscarBravo]

You could always combine remote syslogging with a tool like Splunk to let you sort the wheat from the chaff.

[tog]

Or even use grep which is already there and free on any Linux or BSD box. I have no problems picking out what I need to see.

Even if you're having a problem with 10 clients, concentrate on one, grep for their MAC address in the log, like:
tail -n 2000 logfile | grep 00:01:02:03:04:05

And repeat for other MAC addresses of other customers.

[tkerns]

You are missing the point.... I don't want several AP's sending to syslog hundreds of entries about client dissassociation and association. It seems that every 6 minutes ver 3 will disassociate inactive users, which the majority will re-associate, this results in 2 entries per client every time. After not very long the syslog on the WAR or ver3 becomes over written and any useful info is gone. So what is the point of having a sys log on the WAR ver3 if all the useful info has been replaced by these association messages.

When I am experiencing an issue I go first to the AP, look to see what is going on not my syslog server.

Tim Kerns
CV-Access

[tog]

My logs are divided up. Each AP has its own log file. You're making this harder than it needs to be.

Instead of trying to get the software developer to change the software to suit your whim, why not just do what works reasonably well and accomplishes what you need now? Try it, you might like it.

Use a remote syslog server. Put the logs from each system into separate log files. Use grep (or Splunk or whatever) to look through the log. In fact having grep or splunk available means you can find the relevant information in the log MUCH faster than you could in StarOS's GUI just paging down skimming everything. When I am looking for something or something is going on, I log into the AP to see the current state of everything, and then go to my ssh session with my syslog server to look through some logs to see what has been going on. tail -n 1000 logfile | more to see what has been going on very recently.

I have three types of clients on my network: D-Link G810, WRAP/v2, WAR v3. I have over 20 APs running, I haven't seen this issue you are describing since before beta-16. But, I also do not use CB3s.

Sorry I cannot be of more help, but I really think if you try the remote syslog idea you might find it's faster, easier and ultimately more complete than manually paging through the few logs that StarOS saves that are accessible via its GUI.

If, for example, you are specifically looking for all the logs from the last time your v3 AP booted up, you could try:
tail -n 50000 logfile | more
Then type "/BusyBox" to have it search for the word BusyBox which is contained in the first line that appears at system boot. Then you can press enter to scroll line-by-line from there through all the messages from system boot.

If you use Windows servers only, there are plenty of syslog servers for Windows and you can use search in notepad or whatever rather than grep. Either which way you would get a complete log to look at and you'd have the ability to quickly search for specific things.

[tkerns]

Tog .... back off

I made a suggestion to split the log into 2, if you don't think that is a good idea then say I don't agree instead of trying to tell me how to use a syslog server.

I have one and it collects entries from each AP. I know how to use it.

The syslog on the AP is useless after several minutes because of the entries of associations. The limited space for the log is over written with these entries.

Now, my suggestion is to break it into 2.

All you need to say at this point is I agree or disagree.

Tim Kerns
CV-Access

[tog]

I'm not attacking you, I'm trying to help you. You say that sending things to a remote syslog server is not good enough for various reasons, I try to address the specific points you make to help make it more convenient for you to use the remote syslog server. You tell me to shut up and go away. Forgive me for trying to be overly helpful.

Unless I'm reading you wrong, the whole basis for this thread is that looking at the system info from the system's boot up is somehow useful and you want to forever be able to view that information via the StarOS GUI. Is that correct?

[tony]

While many like the fact that the association, and disassociation messages are shown, it can clog up the on-system log due to it's limited number of lines it shows. I will see about incorporating a method to either split the logs, or at the least, provide an option to mute the association / disassociation messages which may help.

[tkerns]

Thanks Tony,

I do like the association messages and find them useful, but as you say, there are a limited number of lines and other error type messages can become lost.

It is not a big item on my wish list....it was just an observasion.

Tim Kerns
CV-Access