HyperFabric Release Notes (B.11.11.04.02)

HyperFabric Release Notes

Patches and Fixes in This Version

Chapter 1 13

Patches and Fixes in This Version

This section lists the ﬁxes and patches that have been applied in the version B.11.11.04.02 of

the HyperFabric software.

Patches

This release contains no HyperFabric-speciﬁc patches.

Fixes

This release contains the ﬁxes listed in Table 1-3.

Table 1-3 Fixes Included in this Release

CR Number Description

CR JAGae69178 Oracle RAC or HMP point-to-point: Database is not usable in a cluster

with only one active node.

CR JAGae84898 With 5 or more cards, clic_mgmtd daemon dies and dumps core.

CR JAGae80691 During clic_shutdown, ifconfig dumps core.

CR JAGaf21232 Oracle instances could not be spawned with the speciﬁed Value in the

/opt/clic/lib/skgxp/skclic.conf.

CR JAGaf21331 Kernel timer functions could be called with negative timeout values

when Oracle calling in with huge timeout values resulting in adverse

effect on Oracle.

CR JAGaf21249 During failover if a send manages to get into the kernel - if not caught

in the user space - will be returned E3HLT to the SKGXP layer

(proactive change - no reported adverse effects).

CR JAGaf21254 Kernel VCs may be tied to a stale VCs resulting in trafﬁc not

progressing due to long hop routes being replaced by shorter routes.

Oracle could crash if this occurs even when there are no failure events.

CR JAGaf21258 VC ﬁlling logic is changed to be more robust such that less exceptions

are hit in failure path (proactive change - no reported adverse effects).