Bug 968503 - flush_ber error sending back start_tls response will deadlock
Summary: flush_ber error sending back start_tls response will deadlock
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Enterprise Linux 6
Classification: Red Hat
Component: 389-ds-base
Version: 6.4
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: rc
: ---
Assignee: Rich Megginson
QA Contact: Sankar Ramalingam
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-05-29 19:35 UTC by Rich Megginson
Modified: 2018-12-01 15:06 UTC (History)
6 users (show)

Fixed In Version: 389-ds-base-1.2.11.15-22.el6
Doc Type: Bug Fix
Doc Text:
Cause: The server is processing a startTLS request and gets an error (network error usually) attempting to write the response back to the client. Consequence: The server will hang and become unresponsive to this an any new connection request, due to a deadlock in the connection handling code. Fix: Make sure the connection is not locked when this error condition occurs. Result: Server correctly processes startTLS requests when there are network errors, and continues to work fine for established and new connection requests.
Clone Of:
Environment:
Last Closed: 2013-11-21 21:08:19 UTC
Target Upstream Version:

Attachments (Terms of Use)
Add an attachment (proposed patch, testcase, etc.)


Links
System ID Priority Status Summary Last Updated
Red Hat Product Errata RHBA-2013:1653 normal SHIPPED_LIVE 389-ds-base bug fix update 2013-11-20 21:53:19 UTC

Description Rich Megginson 2013-05-29 19:35:51 UTC 
This bug is created as a clone of upstream ticket:
https://fedorahosted.org/389/ticket/47375

If the server is processing a start tls extop, and is sending back the "Start TLS request accepted.Server willing to negotiate SSL." response, and the ber_flush() call in flush_ber() returns non-zero, the server will deadlock.

One problem is that it is very difficult to get the server to return an error at this point.  The customer is using an F5 load balancer and that seems to have something to do with it.
Comment 6 Ján Rusnačko 2013-08-01 12:32:28 UTC 
Is it possible to automate this ? If yes, could you please add verification steps ?
Comment 7 Rich Megginson 2013-08-01 14:18:20 UTC 
(In reply to Ján Rusnačko from comment #6)
> Is it possible to automate this ? If yes, could you please add verification
> steps ?

We do not know how to reproduce this, much less automate.  The customer has verified that the patch works, so I think we will just have to mark the bug as verified by customer.
Comment 9 Ján Rusnačko 2013-09-25 07:32:49 UTC 
Verified sanity only on 389-ds-base-1.2.11.15-24.el6.
Comment 10 errata-xmlrpc 2013-11-21 21:08:19 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHBA-2013-1653.html
Note You need to log in before you can comment on or make changes to this bug.