Opened 7 years ago

Closed 7 years ago

#2570 closed defect (wontfix)

AUTH_DATASRC_CLIENTS_BUILDER_RECONFIGURE_DATASRC_ERROR

Reported by: jreed Owned by:
Priority: medium Milestone:
Component: data source Version:
Keywords: Cc:
CVSS Scoring: Parent Tickets:
Sensitive: no Defect Severity: High
Sub-Project: DNS Feature Depending on Ticket:
Estimated Difficulty: discuss/wontfix? Add Hours to Ticket: 0
Total Hours: 0 Internal?: no

Description

Because of the AUTH_DATASRC_CLIENTS_BUILDER_RECONFIGURE_DATASRC_ERROR issue in #2569, my other datasources were not loaded nor served.

So I removed the broken CH datasource and then on config commit, I got AUTH_DATASRC_CLIENTS_BUILDER_RECONFIGURE_SUCCESS.

b10-auth should be able to serve what it can. A parsing error in one zone shouldn't cause rest of server to be unusable.

The documentation says: "The system is still running with the data sources that were previously configured (i.e. as if the configuration has not changed)." This does not match the behavior as I expect. I think this means if loading a new configuration that fails, but not at startup.

Subtickets

Change History (2)

comment:1 in reply to: ↑ description Changed 7 years ago by jinmei

Replying to jreed:

Because of the AUTH_DATASRC_CLIENTS_BUILDER_RECONFIGURE_DATASRC_ERROR issue in #2569, my other datasources were not loaded nor served.

So I removed the broken CH datasource and then on config commit, I got AUTH_DATASRC_CLIENTS_BUILDER_RECONFIGURE_SUCCESS.

b10-auth should be able to serve what it can. A parsing error in one zone shouldn't cause rest of server to be unusable.

Parsing error in one zone doesn't cause rejection of the entire data
source configuration. It was a known problem and was fixed (hmm, it
should be recorded in ChangeLog).

Note the "Unknown uncaught exception" part of the log message. It
means it was actually a really unexpected situation, not a parser
error. This should be something like a software bug, and the server
could even terminate in that situation. So, the real issue was #2569
itself, not this one.

Can we close this ticket?

comment:2 Changed 7 years ago by shane

  • Milestone New Tasks deleted
  • Resolution set to wontfix
  • Status changed from new to closed

The question is how tolerant of errors we need or want to be.

In the case of a data error, then we should be forgiving. We are, AFAIK.

In the case of a software error, we can try to be forgiving, but basically we need to fix the software.

So, I'm happy if we close this, and fix the bug in #2569.

Note: See TracTickets for help on using tickets.