[error] [sqldb] cannot open database /var/log/flb_kube.db

[andutt]

Bug Report

Describe the bug Openshift 4 CoreOS cri-o fluent-bit v1.8.10 deployed in logging namespace as HOWTO. Role, Rolebinding and serviceaccount created as HOWTO. Deploying daemonset and following errormessage occours.

To Reproduce

• Rubular link if applicable:
• Example log message if applicable:

[1mFluent Bit v1.8.10[0m
* [1m[93mCopyright (C) 2019-2021 The Fluent Bit Authors[0m
* [1m[93mCopyright (C) 2015-2018 Treasure Data[0m
* Fluent Bit is a CNCF sub-project under the umbrella of Fluentd
* https://fluentbit.io

[2022/03/25 09:51:42] [ info] [engine] started (pid=1)
[2022/03/25 09:51:42] [ info] [storage] version=1.1.5, initializing...
[2022/03/25 09:51:42] [ info] [storage] in-memory
[2022/03/25 09:51:42] [ info] [storage] normal synchronization mode, checksum disabled, max_chunks_up=128
[2022/03/25 09:51:42] [ info] [cmetrics] version=0.2.2
[2022/03/25 09:51:42] [error] [sqldb] cannot open database /var/log/flb_kube.db
[2022/03/25 09:51:42] [error] [input:tail:tail.0] could not open/create database
[2022/03/25 09:51:42] [error] Failed initialize input tail.0
[2022/03/25 09:51:42] [error] [lib] backend failed

• Steps to reproduce the problem: Deploy daemonset

Expected behavior Fluent-bit starts and transfer logs to "outside" elasticsearch solution.

Your Environment Openshift 4 CoreOS cri-o fluent-bit v1.8.10

Additional context The only thing changed is Path since pod logs have a different path on CoreOS, like below. Where does fluent-bit want to create this database, is it in the filesystem on the nodes or on a mounted pvc? Since permissions to write on the nodes must likely is prohibited by permissions and security contexts.

[INPUT] Name tail Tag kube.* Path /host/var/log/pods/*.log Parser cri DB /var/log/flb_kube.db Mem_Buf_Limit 5MB Skip_Long_Lines On Refresh_Interval 10

[benz9527]

I also met this question on OpenShift 4.8 CoreOS. I deployed the fluentbit by fluent-operator which developed by kubesphere.