Hi friends,
Running a LAPI server installed with Helm in a K8s cluster. Just short under an hour of running, the pod is restarted. There are no K8s events nor crashloops related to the pod, it just runs and then restarts.
Every time it logs the following:
time="2026-02-13T12:00:58Z" level=warning msg="No matching files for pattern /does/not/exist" module=acquisition.file type=file
time="2026-02-13T12:00:58Z" level=error msg="open /var/lib/crowdsec/data/ip_seo_bots.txt: no such file or directory"
time="2026-02-13T12:00:58Z" level=error msg="open /var/lib/crowdsec/data/rdns_seo_bots.txt: no such file or directory"
time="2026-02-13T12:00:58Z" level=error msg="open /var/lib/crowdsec/data/cloudflare_ip6s.txt: no such file or directory"
time="2026-02-13T12:00:58Z" level=error msg="open /var/lib/crowdsec/data/cloudflare_ips.txt: no such file or directory"
time="2026-02-13T12:00:58Z" level=warning msg="grok 'IPv4_WORKAROUND' already registred" id=summer-violet module=parser name=crowdsecurity/sshd-success-logs stage=s01-parse
time="2026-02-13T12:00:58Z" level=error msg="open /var/lib/crowdsec/data/rdns_seo_bots.regex: no such file or directory"
time="2026-02-13T12:00:58Z" level=warning msg="grok 'IP_WORKAROUND' already registred" id=summer-violet module=parser name=crowdsecurity/sshd-success-logs stage=s01-parse
time="2026-02-13T12:00:56Z" level=warning msg="unable to initialize GeoIP: open /var/lib/crowdsec/data/GeoLite2-City.mmdb: no such file or directory"
time="2026-02-13T12:00:56Z" level=error msg="unable to open GeoLite2-City.mmdb : open /var/lib/crowdsec/data/GeoLite2-City.mmdb: no such file or directory"
level=warning msg="Instance already enrolled. You can use '--overwrite' to force enroll"
time="2026-02-13T12:00:52Z" level=warning msg="Crowdsec service shutting down"
time="2026-02-13T12:00:51Z" level=warning msg="SIGTERM received, shutting down"
According the Helm values example and chart docs, persistent volumes shouldn’t be used in favor of DB storage. But in all examples I find there is persistent volumes configured.
Of course the files aren’t kept over a restart as the pod has no PV, so I need an advice how to configure my helm chart properly. Should I add a PV or not?
Also, is it common for the pods to constantly restart? I really can’t find a reason why they are restarting all the time. No K8s events, no OOM, no error logging except the one logged by the pod.
If I open up a console to the pod and run `cscli hub upgrade` all the missing files are fetched. But still, the pod will restart and everything is back to square one.