Red Hat Bugzilla – Bug 208279
Regression in new traceroute implementation
Last modified: 2007-11-30 17:07:34 EST
RHEL5b1 includes a new traceroute-1.0.4 (epoch 2) implementation.
The main idea of switching to this implementation is to avoid using setuid bit,
i.e. allow to do things even by unprivileged process.
Unfortunately this new implementation has significant regression -- it does not
support tracing using ICMP echo packets (i.e. "traceroute -I").
Tracing of a network using ICMP is very widely used. Absence of such possibility
will create real problems at clients who use RHEL for network infrastructure
(remote servers, routers, firewalls, etc.)
Besides that, the new traceroute's code seems to be a pre-beta stuff, see some
examples in the appropriate Fedora's bug #176587
As is seen in #176587 Dmitry Butskoy developed new traceroute which has all
required features (IPv6 support, ICMP, ICMP6, ...) and same interface as old
traceroute. It uses UDP by default so it does not need suid set. Current version
seems to be stable, but I will do some further testing. I would like to change
to Dmitry's upstream in RHEL5.
This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux release. Product Management has requested further review
of this request by Red Hat Engineering. This request is not yet committed for
inclusion in release.
There is a discussion started in fedora-devel-list, see
QE ack for RHEL5B2 according to section 21a of the release criteria.
The final version (targeted for FC6) is here:
I consider this version as 2.0rc . No more new features for a while.
The final one:
Package is now commited in cvs in REHEL-5 branch and was built successfully.
The latest upstream: