pkg reports 'advisory lock' error

This was a small surprise the other evening. Typing 'pkg install some-thing' reported an error that an 'advisory lock' was held by another process. Checked the process list and found that cron was running a nightly housekeeping script on pkg. After a few minutes had passed, the cron job finished and I was able to resume installing / upgrading packages.

You can see the housekeeping scripts here:

/usr/local/etc/periodic/daily/411.pkg-backup
/usr/local/etc/periodic/daily/490.status-pkg-changes

Good to know the base system does this for you!
 
The daily security emails provide some useful package information. It runs at 3:01am.

For example:
Code:
Checking for packages with security vulnerabilities:
Fetching vuln.xml.bz2: .......... done
python36-3.6.8_2 is vulnerable:
python 3.6 -- multiple vulnerabilities
CVE: CVE-2019-9740
CVE: CVE-2019-9948
WWW: https://vuxml.FreeBSD.org/freebsd/18ed9650-a1d6-11e9-9b17-fcaa147e860e.html

1 problem(s) in 1 installed package(s) found.

Checking for packages with mismatched checksums:

-- End of security output --
It tells me I need to update my Python.
 
My periodic script was running pkg check, from /usr/local/etc/periodic/security/. I did doas ps ax | grep pkg.

I was compiling from ports using portmaster starting before 3am, and there was an error that occurred. When I tried to restart compiling, there was lock. Then, I looked at the time, past 3 am.

I thought the periodic script interrupted my portbuild at first, but the error was about a pkg conflict. And the lock from that build perhaps would prevent the periodic script from running, unless portmaster didn't hold that lock between building different ports.
 
My periodic script was running pkg check, from /usr/local/etc/periodic/security/. I did doas ps ax | grep pkg.

I was compiling from ports using portmaster starting before 3am, and there was an error that occurred. When I tried to restart compiling, there was lock. Then, I looked at the time, past 3 am.

I thought the periodic script interrupted my portbuild at first, but the error was about a pkg conflict. And the lock from that build perhaps would prevent the periodic script from running, unless portmaster didn't hold that lock between building different ports.
I suspect in this situation portmaster ends in error and does not return a locked database error and instead reports something regarding the port is already installed...
 
Back
Top