After Applying Solaris OS Kernal Patch Oracle HTTP Server Stop Responding After a Short Period Of Time
Last updated on JUNE 18, 2018
Applies to:Oracle HTTP Server - Version 18.104.22.168.0 and later
Oracle Solaris on SPARC (64-bit)
After applying Solaris OS patch Oracle HTTP Server (OHS) listener appears to have slow performance and stop responding after running for as short as a few minutes to several hours.
Following symptoms are observed:
- No errors are recorded in log files.
- If multiple listeners are running it is possible that only one listener is affected.
- Issue is not protocol dependent (ssl/non-ssl).
- Sometimes the listener can start responding without any intervention.
- Restarting OHS resolves the issue temporarily.
- 100+ connections in close_wait status. This is probably due to client "GET /" packets ack'd by server side but not received by OHS, thus resulting in client side closing the connection due to
timeout and leaving sockets in CLOSE_WAIT state.
- tcpdumps on OHS server side show issue with network layer, sync and data packets are not ack'd by server/OHS side. Following are two examples.
Sync request not ack'd by server side, remote terminated connection
Issue started after applying Solaris Patch, following are some of the reported Solaris patches
M7 Jan 2018 patchset
OS patch to Oracle Solaris 22.214.171.124.0
Sign In with your My Oracle Support account
Don't have a My Oracle Support account? Click to get started
My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms