Bug 6773 - Missing check for tdbdump in init-script
Summary: Missing check for tdbdump in init-script
Alias: None
Product: CTDB 2.5.x or older
Classification: Unclassified
Component: ctdb (show other bugs)
Version: unspecified
Hardware: x64 Linux
: P3 trivial
Target Milestone: ---
Assignee: Martin Schwenke
QA Contact: Samba QA Contact
Depends on:
Reported: 2009-09-30 09:53 UTC by Paul te Bokkel
Modified: 2016-08-06 10:32 UTC (History)
2 users (show)

See Also:


Note You need to log in before you can comment on or make changes to this bug.
Description Paul te Bokkel 2009-09-30 09:53:31 UTC
On RHEL 5.4 (with RH packaged CTDB 1.0.82), /etc/init.d/ctdb is checking for corrupt tables with tdbdump. However, if it's missing, no warning is given and it's stated that the table is corrupt (which it isn't).

Environment: clean RHEL 5.4 install with CTDB. CTDB was used once, so tables exist. Stop ctdb, try to restart; fail, with message:
Starting ctdbd service: Persistent database /var/ctdb/persistent/account_policy.tdb.1 is corrupted! CTDB will not start.

Install /usr/bin/tdbdump and it works. Trivial to prevent, warn, or add dependency.
Comment 1 Michael Adam 2009-11-20 18:31:55 UTC
Ronnie, I guess we need a RPM dependency to the tdb tools package?
Or to the tdbdump binary?
Comment 2 Michael Adam 2010-02-25 01:41:34 UTC
I have forwarded a patch to Ronnie that adds dependency for /usr/bin/tdbdump and /usr/bin/tdbtool to the spec file. (never versions also require tdbtool.)
Comment 3 Martin Schwenke 2016-08-06 10:32:58 UTC
I fixed this back in 2012 in CTDB commit 00cd75595685dae829758abf1a4cb644af7ed50e.  If a TDB checker can't be found then a warning is logged but startup succeeds.

Since CTDB 2.3, TDB hecking is now done in 00.ctdb event script.