ambra hangs when multiple ambras are in use on the same stack

[russ]

bill reproduced this error when accessing through the webheads, and some research suggests that session binding might be a problem/requirement.

however, on the 0.9.1 final upgrade it seemed like we were able to reproduce this in a situation where session binding could not be the problem (i.e., accessing through different named vhosts, so session binding was ensured by hostname)

[rcave]

This seems to have been fixed by r7462. We aren't experiencing any hangs running multiple Ambra servers in production.

[pradeep]

Replying to rich:

This seems to have been fixed by r7462. We aren't experiencing any hangs running multiple Ambra servers in production.

As I mentioned in #1131, r7462 is not a fix for these kinds of problems. If the cause is unknown and you can't reproduce this any more, please close this as not reproducible or something.

[rcave]

Moving into 0.9.3 for further performance tests.

[rcave]

We observed an Ambra hang in production. Moving into 0.9.3 rc2 for further research.