Description of problem: The IKE daemon creates SAs when needed. Thus, first time a packet comes in, if there isn't an SA, kernel tells IKE to go and establish one. In the meantime, this packet gets dropped. From user or application viewpoint, app has to be issued again since the SA should be established next time around. In LSPP, because more SAs will get created than normally, this can become very problematic. How reproducible: Very. Steps to Reproduce: 1. configure ipsec with racoon 2. do a ping to remote, it will fail. 3. wait a few seconds 4. do the ping again, it should work Actual results: connection: resource temporarily unavailable Expected results: PING x.x.x.x (x.x.x.x) 56(84) bytes of data. 64 bytes from x.x.x.x: icmp_seq=1 ttl=63 time=0.866 ms 64 bytes from x.x.x.x: icmp_seq=2 ttl=63 time=0.742 ms 64 bytes from x.x.x.x: icmp_seq=3 ttl=63 time=0.650 ms
This is fixed upstream now: http://marc.theaimsgroup.com/?l=git-commits-head&m=117104425107990&w=2
per 2/12 discussion, we should build lspp kernel with this fix.
Created attachment 148346 [details] LSPP.65 kernel patch for this issue. This is the upstream backport of this patch. Lots of testing is needed in the LSPP kernel before internal submission is made.
Successfully ran a 16 hour labeled ipsec stress test on the lspp 65 kernel with this patch. So far things look good. I will try a regular ipsec stress test next.
Joy is blaming this patch for the creation of multiple SAs. I will wait to hear results of that discussion before anything else is done here.
There is still some multiple SA discussion. Part of it will be fixed in another kernel patch (IBM has not opened the BZ but will today) and part of it will be handled in the userspace ipsec daemon.
Eric, can we use this bugreport for the change needed in the kernel, the one where we just needed to add the check for the protocol id. That check narrowed it down from triple SAs to double SAs. The userspace patch will hopefully narrow down the double SA to just a single one. So there will be 2 fixes, one for the kernel as described above ( and hopefully can be used for this bug report) and one for userspace for which I can open a new bugreport for. Let me know if this is ok and I will proceed.
No this bug is being used to track the fix so we don't drop the first packet. (see the patch in comment #2) We need a new bug to track the proto matching fix which was not caused by this patch but which was brought to the surface by this patch.
This request was evaluated by Red Hat Kernel Team for inclusion in a Red Hat Enterprise Linux maintenance release, and has moved to bugzilla status POST.
in 2.6.18-16.el5
A fix for this issue should have been included in the packages contained in the RHEL5.1-Snapshot3 on partners.redhat.com. Requested action: Please verify that your issue is fixed as soon as possible to ensure that it is included in this update release. After you (Red Hat Partner) have verified that this issue has been addressed, please perform the following: 1) Change the *status* of this bug to VERIFIED. 2) Add *keyword* of PartnerVerified (leaving the existing keywords unmodified) If this issue is not fixed, please add a comment describing the most recent symptoms of the problem you are having and change the status of the bug to FAILS_QA. More assistance: If you cannot access bugzilla, please reply with a message to Issue Tracker and I will change the status for you. If you need assistance accessing ftp://partners.redhat.com, please contact your Partner Manager.
Verified that this is fixed.
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 the 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-2007-0959.html