Bug 871838
| Summary: | yum updates on RHEL 5 clients fail, checksum issues | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Satellite | Reporter: | Michael Hood <mihood> | ||||
| Component: | Packaging | Assignee: | Katello Bug Bin <katello-bugs> | ||||
| Status: | CLOSED NEXTRELEASE | QA Contact: | Katello QA List <katello-qa-list> | ||||
| Severity: | medium | Docs Contact: | |||||
| Priority: | unspecified | ||||||
| Version: | 6.0.0 | CC: | bkearney, ehelms, jturner | ||||
| Target Milestone: | Unspecified | Keywords: | Triaged | ||||
| Target Release: | Unused | ||||||
| Hardware: | x86_64 | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2014-09-04 18:06:13 UTC | Type: | Bug | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
I've been able to successfully register, subscribe, push and pull content to/from my RHEL5 instances. To further isolate the problem, please ...
1) attach the full 'yum' command entered, and the entire stdout/stderr of that command.
2) include the software versions tested. The following command will provide the desired output: rpm -qa --qf " * %{sourcerpm}\n" "katello*" pulp candlepin
3) run the /usr/share/katello/scripts/katello-debug script on your System Engine and attach the resulting tarball.
Created attachment 640308 [details]
debug
Here is the output from
rpm -qa --qf " * %{sourcerpm}\n" "katello*" pulp candlepin
# rpm -qa --qf " * %{sourcerpm}\n" "katello*" pulp candlepin
* katello-configure-1.1.9-7.el6cf.src.rpm
* katello-qpid-client-key-pair-1.0-1.src.rpm
* pulp-1.1.13-1.el6cf.src.rpm
* katello-1.1.12-16.el6cf.src.rpm
* katello-selinux-1.1.1-1.el6cf.src.rpm
* katello-cli-1.1.8-7.el6cf.src.rpm
* katello-1.1.12-16.el6cf.src.rpm
* katello-candlepin-cert-key-pair-1.0-1.src.rpm
* candlepin-0.7.8.1-1.el6cf.src.rpm
* katello-cli-1.1.8-7.el6cf.src.rpm
* katello-1.1.12-16.el6cf.src.rpm
* katello-qpid-broker-key-pair-1.0-1.src.rpm
* katello-1.1.12-16.el6cf.src.rpm
* katello-1.1.12-16.el6cf.src.rpm
* katello-certs-tools-1.1.8-1.el6cf.src.rpm
(In reply to comment #2) > I've been able to successfully register, subscribe, push and pull content > to/from my RHEL5 instances. To further isolate the problem, please ... > > 1) attach the full 'yum' command entered, and the entire stdout/stderr of > that command. Hey Michael, thanks for the katello-debug attachment and comment#4. I haven't seen this problem when testing RHEL-5.7, 5.8 or 5.9 deployed applications. Can you please still attach the full yum command and output that you see? Typically, yum URL issues tend to be related to the key+cert used to access the repositories on the katello system. I would suggest grabbing the key and cert included in your katello-generated system templates (you can also find the key+cert on registered clients in /etc/yum.repos.d/redhat.repo). Confirm with the curl command that the provided key and cert are correct > curl -k --key katello.key --cert katello.cert https://<fqdn>/pulp/repos/path/to/some/repodata/repomd.xml Looking in the attached production.log, I see some errors related to activities on the system. Could these explain why access to the repositories is denied? > [ERROR: 2012-10-24 15:47:26 #4519] HttpErrors::NotFound: Couldn't find activation key 'bk_inside_key' > [ERROR: 2012-10-25 07:30:24 #4547] HttpErrors::NotFound: Couldn't find activation key 'dec_inside_key' > [ERROR: 2012-10-25 15:57:36 #4547] *** ERROR: Package '0e333892-d03a-4297-be8b-08fbbfdb21a7' not found within the repository (404) *** Asking around in case others have better suggestions/ideas 1) attach the full 'yum' command entered, and the entire stdout/stderr of
> that command.
Well the yum update peice was resolved by installing python-hashlib.
Here is the customers response:
1) yum behaves as expected after I installed python-hashlib.
python-hashlib *might* be in a standard RHEL 5 base install. My virt
guests are based on our court system installations. Those began with
a philosophy of not installing any software that is not needed.
So the client might be a local issue.
I will request the yum peice anyways.
Thanks
Mike
Topic: yum failing on RHEL 5 systems The repo used for the install does not contain python-hashlib. So, as questioned yesterday, the package was never installed (versus installed with @Base and then removed). Here is the list of python packages in that repo (dumped from lynx): [ ] python-2.4.3-46.el5.x86_64.rpm 18-Jan-2012 06:30 58K [ ] python-ctypes-1.0.2-3.el5.x86_64.rpm 15-Dec-2011 07:04 210K [ ] python-devel-2.4.3-46.el5.i386.rpm 18-Jan-2012 06:30 3.0M [ ] python-devel-2.4.3-46.el5.x86_64.rpm 18-Jan-2012 06:30 3.0M [ ] python-dmidecode-3.10.13-1.el5_5.1.x86_64.rpm 06-Sep-2010 07:29 89K [ ] python-docs-2.4.3-1.1.noarch.rpm 18-Jan-2007 13:04 2.5M [ ] python-elementtree-1.2.6-5.x86_64.rpm 18-Jan-2007 13:04 85K [ ] python-ethtool-0.6-5.el5.x86_64.rpm 21-Jun-2011 01:22 30K [ ] python-imaging-1.1.5-7.el5.x86_64.rpm 08-Feb-2011 10:19 406K [ ] python-imaging-devel-1.1.5-7.el5.i386.rpm 08-Feb-2011 10:19 160K [ ] python-imaging-devel-1.1.5-7.el5.x86_64.rpm 08-Feb-2011 10:19 160K [ ] python-iniparse-0.2.3-4.el5.noarch.rpm 04-Mar-2008 02:25 34K [ ] python-lcms-1.18-0.1.beta1.el5_3.2.x86_64.rpm 08-Mar-2009 05:22 609K [ ] python-ldap-2.2.0-2.1.x86_64.rpm 18-Jan-2007 13:04 123K [ ] python-libs-2.4.3-46.el5.x86_64.rpm 18-Jan-2012 06:30 5.9M [ ] python-numeric-23.7-2.2.2.el5_6.1.x86_64.rpm 10-May-2011 03:48 801K [ ] python-pyblock-0.26-4.el5.x86_64.rpm 01-Jul-2009 08:52 61K [ ] python-pycurl-7.15.5.1-8.el5.x86_64.rpm 07-Dec-2010 16:41 73K [ ] python-rhsm-0.98.9-1.el5.noarch.rpm 25-Jan-2012 02:26 66K [ ] python-setuptools-0.6c5-2.el5.noarch.rpm 04-Mar-2008 02:25 480K [ ] python-simplejson-2.0.9-8.el5.x86_64.rpm 25-May-2011 04:03 141K [ ] python-sqlite-1.1.7-1.2.1.x86_64.rpm 18-Jan-2007 13:04 39K [ ] python-suds-0.4.1-2.el5.noarch.rpm 30-May-2011 01:20 251K [ ] python-tools-2.4.3-46.el5.x86_64.rpm 18-Jan-2012 06:30 946K [ ] python-urlgrabber-3.1.0-6.el5.noarch.rpm 29-Nov-2010 02:44 132K I can upload the entire package list in the repo if that would be useful. As is, I'm not sure there is a use for showing the python list. You are correct, python-hashlib is not provided in the base RHEL repository. Please ensure the customer is deploying applications and including the appropriate cf-tools repository. The python-hashlib package is provided in the cf-tools repository. I'm searching the docs to determine where/if we recommend including the companion cf-tools repository. The reason I'm are not seeing this issue in our testing is because we include the 'katello-agent' package in the image template (and target_content.xml). The katello-agent package has the following dependencies ... katello-agent -> gofer -> python-gofer -> python-hashlib In addition, the system templates we use include the base RHEL repository *and* the cf-tools companion repository. These two factors explain why our RHEL5 testing has not exposed this issue. I find it interesting that the yum command was failing without the python-hashlib package installed. Perhaps another package is missing an explicit dependency on python-hashlib? From the yum failure, are you able to determine what code is attempting to use hashlib? Perhaps that might help identify a package that needs an explicit dependency on python-hashlib. This was an issue with a 2.0.X version of pulp. The 2.4 version in Satellite 6 fixes the issue. I am marking this as closed/nextrelease. If you still see this with Satellite 6, please feel free to re-open. |
Description of problem: yum updates on RHEL 5 clients fail, Checksum issues yum is showing checksum problems with all of them. I see no log entries in the clients. On the CFSE server there are error entries in the production.log; e.g., [ERROR: 2012-10-25 08:11:03 #4480] Task package_install (198) is in error state pulp.log has more records, but other than "timeout detected" I don't see anything that hints toward the root cause of the failure. FWIW -- this beta is being run on a local network, CFSE and test virts are all on the same VMWare host, none of the systems are running iptables, only the CFSE host has SELinux enforcing (and ausearch isn't reporting any denials). Version-Release number of selected component (if applicable): How reproducible: I began working with three of them in a group per page 63 of the Evaluation Guide to install an RPM. The package install failed for all three of them -- Result: Request Timeout In case it matters: - These three systems were subscribed using three different activation keys. They all have a common repo that contains the RPM I am attempting to install. - I put one of these systems in a group by itself, attempted the RPM install. It times out, all other failure items (below) are consistent with when I attempted the group with all three systems. Steps to Reproduce: 1. 2. 3. Actual results: the sha256 checksums in the repos produced by CFSE. Expected results: Clients should update fine, without checksum errors. Additional info: Customer resolved this issue by installing python-hashlib-20081119-7.el5.x86_64.rpm