Excessive logging in juju-db
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Won't Fix
|
Undecided
|
Unassigned |
Bug Description
We have an env where juju-db has logged about 40G of messages to /var/log/syslog yesterday.
Version: 1.25.6.1
logging-config:
$ juju get-env logging-config
<root>=WARNING; unit=INFO
I'm seeing a lot of repetition of this in the logs:
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2843] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 511ms
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2782] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 511ms
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2880] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 511ms
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2825] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 511ms
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2728] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 510ms
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2775] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 510ms
Dec 1 15:30:02 state-host mongod.37017[6670]: Thu Dec 1 15:30:02.888 [conn2741] command juju.$cmd command: { getLastError: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:83 510ms
And
Dec 1 06:25:02 state-host mongod.
Dec 1 06:25:02 state-host mongod.
Dec 1 06:25:02 state-host mongod.
Dec 1 06:25:02 state-host mongod.
This has the potential to fill up a disk real fast.
description: | updated |
Juju uses lumberjack library to produce logs. We will be adding compression to the library and modifying both Juju 1.x and Juju 2.x to compress logs.
I am marking this one as a duplicate of bug # 1494661 as we are using it to track this issue.