Confd restarting as Daemon died

CDB and CDB API
1

{“severity”:“debug”,“logger-time”:“2025-02-25T15:10:10.671081”,“program”:“schemaMgr”,“tid”:“0x7de953b7cf00”,“logger”:“SchemaMgrLogger”,“level”:“debug1”,“file”:“schemaMgr.cc”,“method”:“resendMetadataFile”,“line”:1664",“spanId”:“na”,“message”:“something went wrong!”}

2025-02-25T15:10:11.675351+00:00 pod-cfgmgr-fed-upf-scale-upf-1-295d0f2e-agentpool1-md-5fcql-kxhlp confd[315]: devel-cdb inactive client schemMgrCli closed
2025-02-25T15:10:11.675404+00:00 pod-cfgmgr-fed-upf-scale-upf-1-295d0f2e-agentpool1-md-5fcql-kxhlp confd[315]: devel-c Worker closed daemon id: 2, worker id: 18, socket: {socket,<0.342.0>}
2025-02-25T15:10:11.675460+00:00 pod-cfgmgr-fed-upf-scale-upf-1-295d0f2e-agentpool1-md-5fcql-kxhlp confd[315]: - Daemon schemMgrCli died
[2025-02-25T15:10:11][cfgmgrEntryPoint] INFO: Exec container Cfgmgr - Child process termination detected
[2025-02-25T15:10:11][cfgmgrEntryPoint] INFO: Begin container and pod termination.

May I know the reason why schemMgrCli is identified as inactive and that daemon is killed. It is causing confd to restart multiple times.

2

Hi Saranya,

The Community will need more information about the crash in order to help you out. You might paste relevant sections of the logs and if there is a stack trace from the error log in order to give some more context to the situation.

Regards,
Scott

3

See, for example, Inactive client Check_CDB_Phase closed - #4 by cohult

4

Hi @sbarvick, we have assessed this from our side, in this case, ConfigDataMgr is too busy loading schemas and meanwhile the requests received by SchemaMgr are not processed which is causing timeouts and eventually confd pod goes for a restart.
we understand that it could be because the disk speed being too low during the schema load.
we are evaluating it further.

1 Like