From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 Description of problem: Summary: ======== Running "rpm --rebuild" generates the following message: 'error: db4 error(16) from dbenv->remove: Device or resource busy' Background: ============ This bug report was generated at the request of Red Hat support engineer, Paulo Ng. Please refer to Service Request 232865. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Log in as root 2. Run 'rpm --rebuilddb' 3. Actual Results: Relevant output from 'rpm -vv --rebuilddb': =========================================== D: closed db environment /var/lib/rpm/Packages error: db4 error(16) from dbenv->remove: Device or resource busy D: removed db environment /var/lib/rpm/Packages Expected Results: No db4 errors Additional info: Please refer to Service Request 232865. Contact Red Hat support engineer, Paulo Ng. This bug report was created per his request.
Created attachment 91008 [details] Output of 'rpm -vv --rebuilddb'
BTW, running 'cd /var/lib/rpm && /usr/lib/rpm/rpmdb_verify Packages' produces no errors. The Packages database (and every other db in /var/lib/rpm) appears to be fine.
Duplicate of bug 83281
Yup, 'tis the same as 83281. Closing this ticket as duplicate. *** This bug has been marked as a duplicate of 83281 ***
Jeff Johnson closed 83281 without a proper resolution -- Jeff felt too many non-related bugs were being dupped to 83281. This bug (88287) is clearly the same issue as 83281. Further, rpm/db4/nptl issues have not be corrected. Therefore, I am reopening this bug report.
Yes, there is an error message displayed after --rebuilddb. The message is harmless, there is no problem here. This is a known issue and will be fixed in rpm-4.3.
Jeff, even though the error message is harmless, people don't /know/ that it is harmless. I may be the special case, but before I found bug 83281 I thought the message declared something extremely wrong and reinstalled Redhat 9. Shouldn't there be an open bug until the error is fixed and no longer appears? (At the least, this info should be added to the only FAQ I could find about safely rebuilding an RPM database, here: http://rpm.org/hintskinks/repairdb/. I've emailed the editor of the page with the suggestion.)
I understand all that. No matter what, this error message is in already released software, and there are going to be zillions of dupes. All that can't be helped. If it make you happier to have a bug for this problem, that's fine. What I can't do is service a zillion *different* problems in a single bug pile on report, no matter what problem I solve, some other problem will always remain. That's the only rationale for closing the bug, to prevent a catch-all bug pile-on. Feel free to reopen other bugs, or even this bug, and I'll get to the problem when I get to the problem.
> That's the only rationale for closing the bug, > to prevent a catch-all bug pile-on. Feel free > to reopen other bugs, or even this bug, and > I'll get to the problem when I get to the problem. Understood. Personally, I prefer to keep 83281 open. 83281 already has a nice trail... especially with the 'nosysinfo' boot option. Further, although the bug seems to be harmless, the 'nosysinfo' connection indicates a possible awkwardness between rpm and the new nptl implementation, two fairly important components. Therefore, I feel this bug should be treated as a known, unresolved -- albeit harmless -- issue. Anyway, if 83281 is reopened, perhaps we could remove dupped bug reports which do not have clear direct ties to the rpm/db4 issue?