osaf/services/saf/immsv/immnd/ImmModel.cc | 86 +++++++++++++++++++++++++++++-
osaf/services/saf/immsv/immnd/ImmModel.hh | 2 +
osaf/services/saf/immsv/immnd/immnd_init.h | 4 +
osaf/services/saf/immsv/immnd/immnd_proc.c | 43 +++++++++++---
4 files changed, 123 insertions(+), 12 deletions(-)
When discarding a client, there are chances that IMMND fails to send
impl/admo/ccb discard messages or IMMD can't receive those messages.
In that case, we have to make sure that those messages are re-sent.
If not, we will have admo/ccb resource leak or implementers marked as dying
forever.
DEFAULT_TIMEOUT_SEC (6sec) is used as retry interval.
diff --git a/osaf/services/saf/immsv/immnd/ImmModel.cc
b/osaf/services/saf/immsv/immnd/ImmModel.cc
--- a/osaf/services/saf/immsv/immnd/ImmModel.cc
+++ b/osaf/services/saf/immsv/immnd/ImmModel.cc
@@ -534,6 +534,8 @@ static IdVector sImplsDeadDuring
// but before it arrives over
fevs.
// This to avoid apparent
implementor
// re-create by finalizeSync (at
non coord).
+static ContinuationMap2 sDiscardingClientMap; /* Make sure that impl/ccb/admo
associated with
+ the client are discarded
successfully */
static DeferredObjUpdatesMap sDeferredObjUpdatesMap;
@@ -1170,6 +1172,8 @@ immModel_cleanTheBasement(IMMND_CB *cb,
SaUint32T* searchReqArrSize,
SaUint32T** ccbIdArr,
SaUint32T* ccbIdArrSize,
+ SaUint32T** clientIdArr,
+ SaUint32T* clientIdArrSize,
SaUint32T** pbePrtoReqArr,
SaUint32T* pbePrtoReqArrSize,
SaBoolT iAmCoordNow)
@@ -1178,6 +1182,7 @@ immModel_cleanTheBasement(IMMND_CB *cb,
InvocVector searchReqs;
InvocVector::iterator ix1;
IdVector ccbs;
+ IdVector clients;
IdVector pbePrtoReqs;
IdVector::iterator ix2;
unsigned int ix;
@@ -1187,6 +1192,7 @@ immModel_cleanTheBasement(IMMND_CB *cb,
admReqs,
searchReqs,
ccbs,
+ clients,
pbePrtoReqs,
iAmCoordNow);
@@ -1226,6 +1232,16 @@ immModel_cleanTheBasement(IMMND_CB *cb,
osafassert(ix==(*ccbIdArrSize));
}
+ *clientIdArrSize = (SaUint32T) clients.size();
+ if (*clientIdArrSize) {
+ *clientIdArr = (SaUint32T *) malloc((*clientIdArrSize) *
sizeof(SaUint32T));
+
+ for (ix2 = clients.begin(), ix = 0; ix2 != clients.end(); ++ix2, ++ix)
{
+ (*clientIdArr)[ix] = (*ix2);
+ }
+ osafassert(ix == (*clientIdArrSize));
+ }
+
*pbePrtoReqArrSize = (SaUint32T) pbePrtoReqs.size();
if(*pbePrtoReqArrSize) {
*pbePrtoReqArr = (SaUint32T *)
@@ -2146,6 +2162,11 @@ immModel_prepareForSync(IMMND_CB *cb, Sa
ImmModel::instance(&cb->immModel)->prepareForSync(isJoining);
}
+void immModel_addDiscardingClient(IMMND_CB *cb, SaUint32T clientId)
+{
+ ImmModel::instance(&cb->immModel)->addDiscardingClient(clientId);
+}
+
void
immModel_discardContinuations(IMMND_CB *cb, SaUint32T deadConn)
{
@@ -13375,6 +13396,26 @@ ImmModel::discardImplementer(unsigned in
}
void
+ImmModel::addDiscardingClient(SaUint32T clientId)
+{
+ TRACE_ENTER();
+
+ ContinuationMap2::iterator ci = sDiscardingClientMap.find(clientId);
+ if (ci == sDiscardingClientMap.end()) {
+ TRACE("Add discarding client %u", clientId);
+ sDiscardingClientMap[clientId] = ContinuationInfo2(0 /*Unused*/,
DEFAULT_TIMEOUT_SEC);
+ } else {
+ /* Update the create-time */
+ TRACE("Update time for discarding client %u", clientId);
+ osaf_clock_gettime_sec(CLOCK_MONOTONIC, &(ci->second.mCreateTime));
+ osafassert(ci->second.mCreateTime >= ((time_t) 0));
+ }
+
+ TRACE_LEAVE();
+ return;
+}
+
+void
ImmModel::discardContinuations(SaUint32T dead)
{
TRACE_ENTER();
@@ -13730,7 +13771,7 @@ ImmModel::purgeSyncRequest(SaUint32T cli
SaUint32T
ImmModel::cleanTheBasement(InvocVector& admReqs,
- InvocVector& searchReqs, IdVector& ccbs, IdVector& pbePrtoReqs,
+ InvocVector& searchReqs, IdVector& ccbs, IdVector& clients, IdVector&
pbePrtoReqs,
bool iAmCoord)
{
time_t now;
@@ -13875,6 +13916,49 @@ ImmModel::cleanTheBasement(InvocVector&
delete (ccb);
}
+ ci2 = sDiscardingClientMap.begin();
+ while (ci2 != sDiscardingClientMap.end()) {
+ SaUint32T implId;
+ ConnVector cv;
+
+ /* Check if there's still implementer associated with the client */
+ implId = getImplementerId(ci2->first);
+ if (implId) {
+ TRACE("client %llu is not completely discarded", ci2->first);
+ if (now - ci2->second.mCreateTime >= (DEFAULT_TIMEOUT_SEC)) {
+ clients.push_back(ci2->first);
+ }
+ ++ci2;
+ continue;
+ }
+
+ /* Check if there're still CCBs associated with the client */
+ getCcbIdsForOrigCon(ci2->first, cv);
+ if (cv.size()) {
+ TRACE("client %llu is not completely discarded", ci2->first);
+ if (now - ci2->second.mCreateTime >= (DEFAULT_TIMEOUT_SEC)) {
+ clients.push_back(ci2->first);
+ }
+ ++ci2;
+ continue;
+ }
+
+ /* Check if there're still admos associated with the client */
+ cv.clear();
+ getAdminOwnerIdsForCon(ci2->first, cv);
+ if (cv.size()) {
+ TRACE("client %llu is not completely discarded", ci2->first);
+ if (now - ci2->second.mCreateTime >= (DEFAULT_TIMEOUT_SEC)) {
+ clients.push_back(ci2->first);
+ }
+ ++ci2;
+ continue;
+ }
+
+ TRACE("client %llu is completely discarded, remove it", ci2->first);
+ ci2 = sDiscardingClientMap.erase(ci2); /* C++11 */
+ }
+
ci2=sPbeRtReqContinuationMap.begin();
while(ci2!=sPbeRtReqContinuationMap.end()) {
//Timeout on PRT request continuation is hardwired but long.
diff --git a/osaf/services/saf/immsv/immnd/ImmModel.hh
b/osaf/services/saf/immsv/immnd/ImmModel.hh
--- a/osaf/services/saf/immsv/immnd/ImmModel.hh
+++ b/osaf/services/saf/immsv/immnd/ImmModel.hh
@@ -557,6 +557,7 @@ public:
bool reallyDiscard,
IdVector& gv,
bool isAtCoord);
+ void addDiscardingClient(SaUint32T clientId);
void discardContinuations(SaUint32T dead);
void discardNode(unsigned int nodeId, IdVector& cv, IdVector&
gv, bool isAtCoord, bool scAbsence);
void getCcbIdsForOrigCon(SaUint32T dead, IdVector& cv);
@@ -572,6 +573,7 @@ public:
InvocVector& admReqs,
InvocVector& searchReqs,
IdVector& ccbs,
+ IdVector& clients,
IdVector& pbePrtoReqs,
bool iAmCoord);
diff --git a/osaf/services/saf/immsv/immnd/immnd_init.h
b/osaf/services/saf/immsv/immnd/immnd_init.h
--- a/osaf/services/saf/immsv/immnd/immnd_init.h
+++ b/osaf/services/saf/immsv/immnd/immnd_init.h
@@ -133,6 +133,8 @@ extern "C" {
SaUint32T *searchReqArrSize,
SaUint32T **ccbIdArr,
SaUint32T *ccbIdArrSize,
+ SaUint32T **clientIdArr,
+ SaUint32T *clientIdArrSize,
SaUint32T **pbePrtoReqIdArr,
SaUint32T *pbePrtoReqArrSize,
SaBoolT iAmCoordNow);
@@ -379,6 +381,8 @@ extern "C" {
SaUint32T ccbId,
SaUint32T invocation,
SaAisErrorT error, SaUint32T *reqConn);
+ void immModel_addDiscardingClient(IMMND_CB *cb, SaUint32T clientId);
+
void immModel_discardContinuations(IMMND_CB *cb, SaUint32T deadConn);
SaBoolT immModel_immNotWritable(IMMND_CB *cb);
diff --git a/osaf/services/saf/immsv/immnd/immnd_proc.c
b/osaf/services/saf/immsv/immnd/immnd_proc.c
--- a/osaf/services/saf/immsv/immnd/immnd_proc.c
+++ b/osaf/services/saf/immsv/immnd/immnd_proc.c
@@ -133,13 +133,13 @@ uint32_t immnd_proc_imma_discard_connect
if (!scAbsence && immnd_mds_msg_send(cb, NCSMDS_SVC_ID_IMMD,
cb->immd_mdest_id, &send_evt) != NCSCC_RC_SUCCESS) {
if (immnd_is_immd_up(cb)) {
- LOG_ER("Discard implementer failed for
implId:%u "
- "but IMMD is up !? - case not handled.
Client will be orphanded", implId);
+ LOG_WA("Discard implementer failed for
implId:%u "
+ "but IMMD is up !? will retry later ",
implId);
} else {
LOG_WA("Discard implementer failed for
implId:%u "
"(immd_down)- will retry later", implId);
+ cl_node->mIsStale = true;
}
- cl_node->mIsStale = true;
}
/*Discard the local implementer directly and redundantly to
avoid
race conditions using this implementer (ccb's causing abort
upcalls).
@@ -171,14 +171,13 @@ uint32_t immnd_proc_imma_discard_connect
if (immnd_mds_msg_send(cb, NCSMDS_SVC_ID_IMMD,
cb->immd_mdest_id,
&send_evt) != NCSCC_RC_SUCCESS) {
if (immnd_is_immd_up(cb)) {
- LOG_ER("Failure to broadcast discard
Ccb for ccbId:%u "
- "but IMMD is up !? - case not
handled. Client will "
- "be orphanded", implId);
+ LOG_WA("Failure to broadcast discard
Ccb for ccbId:%u "
+ "but IMMD is up !? - will retry
later", implId);
} else {
LOG_WA("Failure to broadcast discard
Ccb for ccbId:%u "
"(immd down)- will retry later",
idArr[ix]);
+ cl_node->mIsStale = true;
}
- cl_node->mIsStale = true;
}
}
free(idArr);
@@ -216,14 +215,13 @@ uint32_t immnd_proc_imma_discard_connect
if(immnd_mds_msg_send(cb, NCSMDS_SVC_ID_IMMD,
cb->immd_mdest_id,
&send_evt) != NCSCC_RC_SUCCESS) {
if (immnd_is_immd_up(cb)) {
- LOG_ER("Failure to broadcast
discard admo0wner for ccbId:%u "
- "but IMMD is up !? -
case not handled. Client will "
- "be orphanded", implId);
+ LOG_WA("Failure to broadcast
discard admo0wner for ccbId:%u "
+ "but IMMD is up !? -
will retry later", implId);
} else {
LOG_WA("Failure to broadcast
discard admowner for id:%u "
"(immd down)- will
retry later", idArr[ix]);
+ cl_node->mIsStale = true;
}
- cl_node->mIsStale = true;
}
}
}
@@ -234,6 +232,12 @@ uint32_t immnd_proc_imma_discard_connect
if (!cl_node->mIsStale) {
TRACE_5("Discard connection id:%llx succeeded",
cl_node->imm_app_hdl);
+
+ /* Add the discarding client to continuation map */
+ immModel_addDiscardingClient(cb, client_id);
+
+ /* Don't add stale clients to the map. They will be discarded by
+ * immnd_proc_imma_discard_stales() when IMMD is up again. */
}
TRACE_LEAVE();
@@ -986,6 +990,8 @@ static void immnd_cleanTheHouse(IMMND_CB
SaUint32T searchReqArrSize = 0;
SaUint32T *ccbIdArr = NULL;
SaUint32T ccbIdArrSize = 0;
+ SaUint32T *clientIdArr = NULL;
+ SaUint32T clientIdArrSize = 0;
SaUint32T *pbePrtoReqArr = NULL;
SaUint32T pbePrtoReqArrSize = 0;
@@ -1082,6 +1088,8 @@ static void immnd_cleanTheHouse(IMMND_CB
&searchReqArrSize,
&ccbIdArr,
&ccbIdArrSize,
+ &clientIdArr,
+ &clientIdArrSize,
&pbePrtoReqArr,
&pbePrtoReqArrSize,
iAmCoordNow);
@@ -1261,6 +1269,19 @@ static void immnd_cleanTheHouse(IMMND_CB
free(ccbIdArr);
}
+ if (clientIdArrSize) {
+ for (ix = 0; ix < clientIdArrSize; ++ix) {
+ /* The client-node is deleted in the first attempt of
discarding.
+ * Create a dummy client-node for the client-id */
+ IMMND_IMM_CLIENT_NODE dummy_node;
+ memset(&dummy_node, 0, sizeof(IMMND_IMM_CLIENT_NODE));
+ dummy_node.imm_app_hdl =
m_IMMSV_PACK_HANDLE(clientIdArr[ix], cb->node_id);
+
+ immnd_proc_imma_discard_connection(cb, &dummy_node,
false);
+ }
+ free(clientIdArr);
+ }
+
if (pbePrtoReqArrSize) {
IMMSV_EVT reply;
memset(&reply, 0, sizeof(IMMSV_EVT));
------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Opensaf-devel mailing list
Opensaf-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensaf-devel
