Bug 24673
| Summary: | pump -d still updates resolv.conf | ||
|---|---|---|---|
| Product: | [Retired] Red Hat Linux | Reporter: | Wade Minter <minter> |
| Component: | pump | Assignee: | Elliot Lee <sopwith> |
| Status: | CLOSED WONTFIX | QA Contact: | David Lawrence <dkl> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 7.1 | CC: | saint |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2001-07-16 13:30:34 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Wade Minter
2001-01-23 04:36:27 UTC
This defect is considered MUST-FIX for Florence Release-Candidate #1 This works fine for me -- can you let me know exactly what you're seeing, and how pump is getting run. I also just checked pump operation with pump-0.8.7-1. "pump -s -i eth0" on a system with working interface and valid lease wiped out my /etc/resolv.conf. This is very interesting. The problem is that pump runs as a background daemon. Once it is first started on an interface, it will remain running (by default) as long as the interface remains up. It does this, so it can do lease renewals, etc. When pump is first started, it accepts command line parameters, as well as settings from the pump config file (/etc/pump.conf unless set with -c). If you enter -d on the command line when you first start pump for an interface, -d _DOES_ work. It works so well that any subsequent command-line initiated pump's (e.g. a subsequent "pump -i eth0") will also not touch the resolv.conf file. The reverse is also true. I start pump without -d as a parameter on an interface, e.g. eth0. Subsequent "pump -d -i eth0" will cause pump to do its thing, but the -d option will have no effect. Summarised, the "overrides" options on the command line only have any effect if specified for the first pump instance on a given interface. Why? If we specify -d for example on a subsequent command line, it gets parsed and overrides->flags is set to 1. When we write to the control socket cont, we do not pass overrides->flags at all. On the receiving end, in runDaemon, we refer to overrides->flags, however this is using the overrides that was passed to the function. This needs to be cleaned up. Reality doesn't match the man page and command syntax at the moment. I'll try and do this post-florence Given that dhcpcd replaced pump in rawhide, and that the use case seems very esoteric, just going to wontfix this. Disagreement welcome. Apologies for the unresponsiveness of the previous pump packager... |