Bug 589636
Summary: | rpmbuild cycles during test part when ipv6 is enabled | ||||||
---|---|---|---|---|---|---|---|
Product: | Red Hat Enterprise Linux 6 | Reporter: | Aleš Mareček <amarecek> | ||||
Component: | nss | Assignee: | Elio Maldonado Batiz <emaldona> | ||||
Status: | CLOSED ERRATA | QA Contact: | Aleš Mareček <amarecek> | ||||
Severity: | high | Docs Contact: | |||||
Priority: | high | ||||||
Version: | 6.0 | CC: | azelinka, borgan, ddumas, dgregor, ebenes, jrieden, rrelyea | ||||
Target Milestone: | rc | Keywords: | Reopened, TestBlocker | ||||
Target Release: | --- | ||||||
Hardware: | All | ||||||
OS: | Linux | ||||||
Whiteboard: | |||||||
Fixed In Version: | nss-3.12.9-4.el6 | Doc Type: | Bug Fix | ||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2011-05-19 14:03:22 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: | |||||||
Bug Depends On: | |||||||
Bug Blocks: | 580448 | ||||||
Attachments: |
|
Description
Aleš Mareček
2010-05-06 15:25:51 UTC
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux major release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Major release. This request is not yet committed for inclusion. This problem seems to be the same as that reported in Bug 539183 which is now closed. It certainly doesn't happen with the brew builds. I wonder if it may be due to a particular network configuration setting in the machine in question. Can you consistently reproduce it? 1. Reserve machine in RHTS, RHEL6.0-Snapshot-2 2. Now, you can connect via ssh (there is an IP on eth0 and lo0) 3. Download the latest nss srpm for RHEL6 and try to reuild it 4. you should get cycle in test part of srpm Additional info: localhost and localhost.localdomain answered by 127.0.0.1 (ping) localhost and localhost.localdomain answered by ::1 (ping6) ### LOG ### tstclnt: connecting to localhost.localdomain:9432 (address=::1) tstclnt: connect: Operation is still in progress (probably a non-blocking connect). tstclnt: about to call PR_Poll for connect completion! tstclnt: PR_Poll returned 0x28 for socket out_flags. tstclnt: unable to connect (poll): Connection refused by peer. ########### In log, as yoou can see, is connection to IPv6 only or only this test is cycling? (In reply to comment #5) > What status is here please? Some news? I cannot reproduce the problem. I was able to do a rebuild of nss-3.12.6-3.el6.i686 in a system running RHEL 6 beta 2 updated to the latest fom the nightly internal repository. I downloaded the srpm from brew and executed rpmbuild --rebuild nss-3.12.6-3.el6.src.rpm as per your instuctions and it completed just fine. No changes to nss were required and I did this in a slow machine. Please verify that you can rebuild nss-3.12.6-3 with the most recent RHEL 6, I used Today's nighlt wich should be the same as Snapshot 7. Since comment 9 describes a workaround, and given where we are in RHEL6, we should move this to 6.1. Do we want a release note with the workaround? By the way, I am trying to comment out lines in the spec file incorrecly. #%{...} doesn't work, it should be #%%{..} or something else. Getting back to the problem at hand, I can't reproduce it on my sytem of in the build machines. I enabled all the test suites and had no hang-ups in my system. Submitted a scratch build https://brewweb.devel.redhat.com/taskinfo?taskID=2704937 As I write several have completed without problems. Nss builds fail under root user due to a failure in following two tests from testsuite executed during the build: dbtests.sh: #204: Dbtest r/w succeeded in an readonly directory 0 - FAILED dbtests.sh: #4213: Dbtest r/w succeeded in an readonly directory 0 - FAILED The same problem has been reported in Fedora as Bug 646045. According to Elio's analysis these two tests are for non-root users. I've done rebuild under non-root user and all tests passed. One of the ways for reparation could be create a user in test and then use it for these two tests. Other way could skip these tests while rebuilding under root but I would not recommend this "fix". Ales's workaround is fine for now. I have a patch for the tests script at https://bugzilla.redhat.com/attachment.cgi?id=455374&action=diff that allows the test to be run as root. That will have to wait for RHEL 6.1 or RHEL 6.0_Z. (In reply to comment #35) Even with the patch we should still find a way to create a non-root user to run the tests as Ales states, otherwise we aren't getting the intended testing. Hi Elio, I'm reopening this bug. Now, we're testing erratas' TPS and this nss test breaks all system like no tps job can continue because it freezes (cycles) on nss rebuild. I think it's not a proper behaviour that test can cycle and nothing can stop it. Anyway, what does it mean "kill -0" ? See description log: kill -0 26327 >/dev/null 2>/dev/null In case of /etc/hosts "fix": What I discovered is that /etc/hosts like following is correct so I can't change it forever, it wouldn't be right solution. $ cat /etc/hosts 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4 ::1 localhost localhost.localdomain localhost6 localhost6.localdomain6 The problem has been independently reported upstream. A patch to the selverv test tool has been suggested as a temporary fix for Linux that would allow the ssl tests to proceed under ipv6 and I am willing to apply it as soon as it's available. See https://bugzilla.mozilla.org/show_bug.cgi?id=617723#c13 Until we get a response from upstream, I'll pick he patch from https://bugzilla.mozilla.org/attachment.cgi?id=499383&action=diff as a temporary workaround so we unblock testing for the beta compose. This patch which works for Fedora breaks the RHEL 6.1 builds. It seems that ipv6 is not enabled in the build system. We will have to wait for the better temporary fix that was mentioned in the upsstream bug report. I will contact our upstream friends on the issue. Created attachment 480793 [details]
Short-term fix for ssl test suites hangs on ipv6 type connections
Better patch from upstream. Change selfserv to use a dual-stack IPv6 listening socket, which can accept connections from both IPv4 and IPv6 clients. NSPR's IPv6 sockets have the IPV6_V6ONLY socket option default to false. My tests look good so far. Waiting on results from other platforms before I apply it. There is a brew scratch of nss-3.12.9-4.el6 in case someone else wants to test it.
Comment on attachment 480793 [details]
Short-term fix for ssl test suites hangs on ipv6 type connections
r+ by the NSS team.
Checking in nss-589636.patch;
/cvs/dist/rpms/nss/RHEL-6/nss-589636.patch,v <-- nss-589636.patch
initial revision: 1.1
Checking in nss.spec;
/cvs/dist/rpms/nss/RHEL-6/nss.spec,v <-- nss.spec
new revision: 1.68; previous revision: 1.67
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2011-0692.html |