Scribe Scribe v3.2.5 [2025-Nov-23] - Entware syslog-ng and logrotate installer for Asuswrt-Merlin

[elorimer]

Yep, the message has already been modified accordingly.

That line in the .asp must set a record for length!

If I follow, this message will show if the timeout is triggered. That would also happen if the file is too big (takes too long) to load. And, can the log file ever not exist for uiScribe? I know syslog-ng won't create the file if it doesn't exist until it wants to write to the destination file.

One of the problems with scribe is its dependence on logrotate being configured exactly right. You have to have a logrotate.d configuration file for every log file. Otherwise, the overall maxsize of 4mb contained in A00global will trigger a rotation of the log but no restart of syslog-ng, and no logging at all. I've been caught out with that error a number of times, and your implementation of showing the size of a log file is going to be a huge help. I have been noodling, though, on the idea of implementing syslog-ng's native rotation, with a setting of

file("/var/log/my-logfile.log" logrotate(enable(yes), size(4MB), rotations(5)));

which would need to be in every syslog.d destination. But that would eliminate the problem of a non-existent logrotate.d file, the problem of syslog-ng hanging, and the problem of any log file becoming too big, the need to adjust the cron job, and maybe some other user errors. Dunno.

 

[Martinski]

That line in the .asp must set a record for length!

LOL!!

... That would also happen if the file is too big (takes too long) to load.

Nope, that's incorrect. There's a separate message for the case where the file cannot be loaded for some reason (e.g. being too large - well over 4MB).

And, can the log file ever not exist for uiScribe?

It's unusual, but it can certainly happen. A friend of mine was running the BETA version last night, and right after the initial installation, he found that the wlceventd.log log file did not exist for some time.

I know syslog-ng won't create the file if it doesn't exist until it wants to write to the destination file.

Yep, another case in point.

One of the problems with scribe is its dependence on logrotate being configured exactly right. You have to have a logrotate.d configuration file for every log file. Otherwise, the overall maxsize of 4mb contained in A00global will trigger a rotation of the log but no restart of syslog-ng, and no logging at all. I've been caught out with that error a number of times, and your implementation of showing the size of a log file is going to be a huge help. I have been noodling, though, on the idea of implementing syslog-ng's native rotation, with a setting of

file("/var/log/my-logfile.log" logrotate(enable(yes), size(4MB), rotations(5)));

which would need to be in every syslog.d destination. But that would eliminate the problem of a non-existent logrotate.d file, the problem of syslog-ng hanging, and the problem of any log file becoming too big, the need to adjust the cron job, and maybe some other user errors. Dunno.

Some people would say the combination of the logging utility plus the log rotating tool makes it more powerful because you can fine-tune the combo to meet your own specific needs and requirements, based on your system resources, the volume of log entries being generated on a daily basis, and the purpose of keeping more than 3MB to 4MB worth of log files - especially if you're not a system admin doing weekly reviews and historical analysis of the system logs looking for any signs of system malfunctions, external intrusions, unusual resource consumption, etc.

OTOH, I do understand that using those tools properly and to their full potential requires users to be a bit more technically savvy about their system and Linux in general. It's really a double-edged sword.

I'm still relatively new to scribe and logrotate myself and still learning the ins and outs, so I can't really help you with the details of all possible configurations.

Just my 2 cents.

 

[elorimer]

Well, at current exchange rates the 2 cents of the maintainer are worth more than the 2 cents of this here gawker.

I just harbor the suspicion that no one is paying the slightest attention to the logrotate options. Not minsize, not compression, not the rotate number and cetainly not daily. 10 log files, max 4MB, with 4 prior versions, is only 200MB total. And folks actively scrutinizing the log files would be better sending them to one of the services with more analytics. We fooled a bit with loggly once.

 

[Tarek Yag]

Holy Smoke, Batman!! You seem to want "credit" for posting or even being on this forum!!

LOL, I'm not slipping down again into another argument with such a rude and disrespectful person. Especially knowing that you have the higher hand, and people usually cheer bosses rather than rights! That's just pathetic and such a shame.

it was agreed that you would submit a PR with your solution so we could review it for integration.

Well, after the last few incidents, it's obvious that I shouldn't invest any more time/effort into a PR (other than convincing you and explaining the feature in FULL details) to not be "credited" -the word that you, and some others, hate- in the end!

I'll take your reply as a rude admission that you used my proposed idea in details, and the idea from @penguin22 for log rotation:

It would be very helpful to have the ability to clear specific logs or logrotate from the uiScribe GUI. I was hunting for this capability, which appears to only be without Scribe installed and had to go hunting through SSH to get it cleared for the router to be responsive again.

Crediting it to "some people" as if they were too many and hard to mention! I feel just sorry for such mentalities!

worth more than the 2 cents of this here gawker

LOL, congrats for your 2 cents! Hoping that you find another two specific issues with Scribe for the benefit of everyone! That gawker already acknowledges more bugs and flaws within Scribe, he's just glad now that he didn't and won't report them.

I won't comment anymore in this regard.

 

[Martinski]

Well, at current exchange rates the 2 cents of the maintainer are worth more than the 2 cents of this here gawker.

I just harbor the suspicion that no one is paying the slightest attention to the logrotate options. Not minsize, not compression, not the rotate number and cetainly not daily. 10 log files, max 4MB, with 4 prior versions, is only 200MB total. And folks actively scrutinizing the log files would be better sending them to one of the services with more analytics. We fooled a bit with loggly once.

You're probably correct; the vast majority of users are likely leaving the default options & settings assigned in the configuration files and don't check other possible options. For ASUS routers, though, I can only assume that @cmkelley, with the help of @Jack Yaz and other forum users who participated in testing and troubleshooting the early development and production releases, arrived at those default settings after some experimentation and consideration of how most users configure their home routers.

 

[Davidncali001]

The screenshots and output shown in your latest post look correct, and they indicate that your Scribe setup is currently configured as expected. If your JFFS partition continues to get filled up again, I suspect it might be caused by something else, but perhaps it leads to a corruption in the Scribe functionality.

I have a custom shell script that I've used before as a diagnostic tool to monitor cases when either the TMPFS or the JFFS file system was being filled up slowly with large files over a period of a few weeks. IIRC, the last time was about 2 years ago, where the Traffic Monitor (or Traffic Analyzer??) was generating a very large database file (a little over 40MB). The log file generated by the diagnostic script and stored in the USB-attached drive showed the database file slowly increasing and filling the JFFS partition during the previous 3 weeks.

In your particular case, I don't really know what's causing the problem, but you could set up the diagnostic script to run as a cron job at a fixed interval (e.g. every 20 mins or every 4 hours), depending on how fast the JFFS is getting filled up and, hopefully, the log file will capture some rogue file(s) getting larger over time.

You can use the following commands to download the custom script from my personal GitHub repo:

curl  -LSs --retry 3 --retry-delay 5 --retry-connrefused \
https://raw.githubusercontent.com/Martinski4GitHub/CustomMiscUtils/master/Diags/LogMemoryStats.sh \
-o /jffs/scripts/LogMemoryStats.sh && chmod a+x /jffs/scripts/LogMemoryStats.sh

Once downloaded into the router, type the following command:

/jffs/scripts/LogMemoryStats.sh -help

The output will provide some useful CLI syntax to set up the cron job and other parameters to configure the diagnostic script for your own specific needs. Try to run the script by itself first to see the kind of output that gets generated, and if you're willing to let it run as a cron job, execute the call to set it up with your preferred time interval. If you want the cron job to persist across reboots, copy & paste the given command at the end of your '/jffs/scripts/post-mount' hook script.

Example call:

/jffs/scripts/LogMemoryStats.sh -setcronjob 1hour

HTH.

thank you, i will download and run the script. Really appreciate your time and efforts.

David

 

[Martinski]

Release Notes for Scribe v3.2.6_25120600 BETA-2 development version now available
[2025-Dec-06]

1) NEW: Added a new 'A01global' configuration file to set more default global directives for log rotation.

These default global directives take effect unless overridden by options set in a separate configuration file associated with a specific log file (e.g. "/opt/etc/logrotate.d/messages").​

2) IMPROVED: Cron job to rotate logs was modified. All calls to logrotate will now include filtered log files that don't have a corresponding configuration file in the "/opt/etc/logrotate.d" directory.

This means that users don't need to add a logrotate configuration file for each filtered log file unless they want to apply additional or different options/directives from the global defaults. Upon a fresh installation of Scribe, the only individual configuration files included in the "/opt/etc/logrotate.d" directory by default are messages, logrotate, and syslogng in addition to the global files (A00global, A01global).​

3) IMPROVED: Added a mutually exclusive lock for the calls to logrotate made from Scribe and uiScribe to prevent a possible case where simultaneous executions will attempt to rotate the same set of log files.

IMPORTANT NOTE:
Installation of this latest BETA development version of Scribe is required for any future uiScribe BETA development versions.

To get the latest BETA development v3.2.6_25120600 version, run the following commands:
<CODE>
/jffs/scripts/scribe develop
/jffs/scripts/scribe forceupdate
<CODE>

P.S.
I ran out of time to test and validate the latest changes of uiScribe before making the next BETA-2 release, so I'll do that when I get some free time over the weekend.

 

[kernol]

Beta-2-dev running sweetly on my GT-AX6000.

After a full reset and rebuild from scratch a while back I had forgotten to add a logrotate file for my custom "notmesh" syslog-ng filter ... had to uninstall and reinstall after runaway on my webui access to System Log tab.

At least now - when I forget next time - your A01Global will be there as a catch-all - thanks