Red Hat Bugzilla – Bug 804156
libssh2 fails key re-exchange when write channel is saturated
Last modified: 2012-03-21 14:43:03 EDT
+++ This bug was initially created as a clone of Bug #804145 +++
Description of problem:
This bug is fixed upstream:
It's a blocker for virt-p2v. virt-p2v sends a large amount of data over ssh using libssh2. When it has sent enough data (about 64G), the peer will request key re-exchange. Because virt-p2v sends data continuously, it will hit this bug with almost 100% reliability, making it impossible to transfer more than 64G of data.
Version-Release number of selected component (if applicable):
--- Additional comment from email@example.com on 2012-03-16 12:08:23 EDT ---
Created attachment 570649 [details]
Patch to openssh to aid testing
It takes about 45-60 mins of sending data on my system to hit this bug. I'm using a version of openssh with this patch to aid testing. It causes openssh to force a rekey 32x sooner than normal.
--- Additional comment from firstname.lastname@example.org on 2012-03-16 12:43:26 EDT ---
Created attachment 570652 [details]
Backport of upstream fix
This is a backport of upstream commit cc4f9d5679278ce41cd5480fab3f5e71dba163ed. The backport was uncomplicated: code has shifted a bit, but nothing close to the affected lines has been altered.
--- Additional comment from email@example.com on 2012-03-16 12:54:28 EDT ---
Created attachment 570653 [details]
Backport of upstream fix (V2)
Spoke too soon! libssh2_kex_exchange was renamed to _libssh2_kex_exchange, so the above doesn't compile. This version is fixed.
libssh2-1.4.0-2.fc17 has been submitted as an update for Fedora 17.
* should fix your issue,
* was pushed to the Fedora 17 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing libssh2-1.4.0-2.fc17'
as soon as you are able to.
Please go to the following url:
then log in and leave karma (feedback).
libssh2-1.4.0-2.fc17 has been pushed to the Fedora 17 stable repository. If problems still persist, please make note of it in this bug report.