search

google-code-export / lusca-cache

Automatically exported from code.google.com/p/lusca-cache
0 stars 0 forks source link

URGENT: random memory corruption in LUSCA_HEAD #114

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
I'm seeing random memory corruption on one of the busy cacheboy edge nodes.

This was stable running an earlier version of Lusca.

Let's start tracing down exactly which commit(s) introduced instability and
figure it out.

  * r14705: unstable
  *r14635:

Original issue reported on code.google.com by adrian.c...@gmail.com on 30 May 2010 at 2:31

GoogleCodeExporter commented 9 years ago 
r14635 lasted a good 12 hours with no issues. Going back to r14705 to see if the
crashes return.

Original comment by adrian.c...@gmail.com on 30 May 2010 at 10:27

GoogleCodeExporter commented 9 years ago 
Crashes returned. Trying r14674 next.

Original comment by adrian.c...@gmail.com on 30 May 2010 at 11:00

GoogleCodeExporter commented 9 years ago 
r14674 crashed. The crash is between 14635 and 14674. Going to try r14655 next.

Original comment by adrian.c...@gmail.com on 30 May 2010 at 12:22

GoogleCodeExporter commented 9 years ago 
r14655 seems fine. The crash is between 14655 and 14674.

Trying 14665 next.

Original comment by adrian.c...@gmail.com on 31 May 2010 at 4:40

GoogleCodeExporter commented 9 years ago 
Something that showed up in 14655 (and thus may be around in any recent 
version) :

2010/05/30 06:42:50| WARNING: Unparseable content-length '21777103'

.. that looks fine. Why's it unparsable on this box? Add a test case?

Original comment by adrian.c...@gmail.com on 31 May 2010 at 4:43

GoogleCodeExporter commented 9 years ago 
14655 is fine. This is getting very close to being the actual HTTP vector 
commit.
What next..

Original comment by adrian.c...@gmail.com on 31 May 2010 at 1:31

GoogleCodeExporter commented 9 years ago 
r14671 seems stable. So it's 14672->14674.

Original comment by adrian.c...@gmail.com on 1 Jun 2010 at 6:09

GoogleCodeExporter commented 9 years ago 
r14672 is seeing the random crashes. I'll leave it be for a while; then I'll try
including some hacks to bzero() the HttpHeaderEntry before it's used. It may be 
left
over cruft in the entry confusing other code.

Original comment by adrian.c...@gmail.com on 1 Jun 2010 at 11:36

GoogleCodeExporter commented 9 years ago 
Those bzero() changes didn't stop the crashes. I'll revert and then analyse the
patch. See what else can be done.

Original comment by adrian.c...@gmail.com on 1 Jun 2010 at 3:05

GoogleCodeExporter commented 9 years ago 
The node is back to r14671 and is again stable.

Original comment by adrian.c...@gmail.com on 2 Jun 2010 at 5:52

GoogleCodeExporter commented 9 years ago 
I've pulled that vector patch out of LUSCA_HEAD. It's now stable.

Question is why it introduced the issue..

Original comment by adrian.c...@gmail.com on 23 Jun 2010 at 3:29