Rescuing a Linux system from close to catastrophe

By | August 13, 2021

The extra you understand about how Linux works, the higher you will find a way do some good troubleshooting whenever you run into an issue. On this publish, we’ll dive into an issue {that a} contact of mine, Chris Husted, lately bumped into and what he did to find out what was taking place on his system, cease the issue in its tracks, and be sure that it was by no means going to occur once more.

It began when Chris’ laptop computer reported that it was working out of disk space–specifically that solely 1GB of accessible disk area remained on his 1TB drive. He hadn’t seen this coming. He additionally discovered himself unable to save lots of information and in a really difficult state of affairs since it’s the solely system he has at his disposal and he wants the system to get his work achieved.

When he was prompted by the system to “Look at or Ignore” the issue, he selected to look at it. Trying round, he seen that his /var/log listing had turn into extraordinarily giant. Inspecting the listing extra intently, he noticed that his syslog file had grown to 365GB. Think about being Chris and one thing like this:

Searching round on the internet, Chris discovered this publish on stackoverflow that inspired capping the scale of the syslog file.



The very first thing he did was run these three instructions:

The primary command allowed him to tackle root privileges, the second emptied the syslog file on the system and the third restarted the syslog daemon so it could proceed to gather details about what was taking place on the system. He nonetheless wanted to trace down the perpetrator.

Subsequent, he modified his logrotate settings (within the /and many others/logrotate.d/syslog file) so the file couldn’t turn into any bigger than 1GB. He did this by including the maxsize setting as identified within the strains under:

The primary line (rotate 7) ensures that seven generations of the syslog file can be retained together with the present one, however does not resolve issues by which the present file grows to an infinite measurement in a single day. On a traditional system, the gathering of syslog information will look one thing like this when rotated each day:

The mixture of “rotate 7” (maintain seven generations) and “each day” (rotate every single day) leaves you with a set of information like these proven. Including the maxsize setting implies that your logs will rotate each day or each time they attain the scale specified, so that you is perhaps rotating logs greater than as soon as a day. Given the 1G setting, nevertheless, you must by no means see the information utilizing greater than 1GB for the present and former information and sure lower than a tenth that measurement for the rest of the logs since they will be compressed. This can be sure that the syslog information will not seemingly use greater than 3 GB in total–far smaller than Chris’ 365 GB. (You may get extra element on how log rotation works from this publish.)

With the scale of the syslog file constrained, Chris was able to delve into the reason for the issue. First, he ran this command:

This allowed him give attention to the underside of the file, but additionally displayed further strains as they had been being added. A stream of messages together with strings like “baloo_file.desktop[2982]: org.kde.baloo.engine:” shortly recognized Baloo (the file indexing and file search framework for KDE Plasma) because the supply of the issue.

Since Chris was utilizing Ubuntu GNOME, he wanted to look into why Baloo was working on his system in any respect. Then he recalled he had put in a file supervisor named Dolphin that may have introduced Baloo together with it.

Utilizing the balooctl command, he was in a position to confirm that baloo was certainly working and stopped it utilizing these instructions as root:

Then he eliminated Dolphin (which Software program Supervisor hadn’t helped with) utilizing these instructions:

Afterwards, Chris’ system was instantly again on top of things, and he had recovered 300GB of his disk area. After a bit extra home cleansing, clearing caches, eradicating no-longer-used apps, and many others., Chris had recovered greater than 400GB of drive area. He claims that now hislaptop runs as quick because it did when Ubuntu was first put in.

Notice that some Linux techniques use messages information as an alternative of syslog information, and that others (like Fedora) now use the journalctl command to show knowledge saved in information saved within the /var/log/journal listing.

A worrisome drawback and one which made a Linux laptop computer nearly fully unusable was resolved with good perception on find out how to liberate some disk area and cease the disk from filling up, a fast evaluation of the issue by reviewing the syslog file entries, a modification of log rotation settings and eradicating the system companies that had been inflicting the issue.

I ought to emphasize that Chris considers himself a Linux consumer, not a “techie”, and was grateful to trace down and repair the issue himself with freely out there assist from different Linux customers or, as Chris describes it, “real experience defined in plain English for common individuals”. He harassed how vital that is for him as a Linux consumer and the way vital he imagines that is for all of us.

Given Chris’ expertise, possibly extra of us ought to think about capping the scale of our log information, monitoring disk-space utilization, and by no means forgetting how a lot assist is offered for us on-line.