2779420 - HANA crashes if using MAX_RECOVERY_BACKINT_CHANNELS=64

If use more than one (until 64) MAX_RECOVERY_BACKINT_CHANNELS in HANA 1.0 or HANA 2.0, a log restore from tape or disk hang until cancelled or timeout occurs.
If changing the parameter from 64 to 1, the recovery works fine again, but need long time for recovery of big Databases.


Backup.log:
2018-12-14T09:14:25+01:00 P005419 167abc5b20e INFO RECOVERY state of service: indexserver, abc-d-efg001:30040, volume: 2, RecoveryLogReplayStarted
2018-12-14T09:14:25+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 0% 0/76
2018-12-14T09:14:34+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 0% 0/9
2018-12-14T09:14:34+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 11% 1/9
2018-12-14T09:14:34+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: xsengine, abc-d-efg001:30043, volume: 3, 50% 1/2
2018-12-14T09:14:34+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: xsengine, abc-d-efg001:30043, volume: 3, 100% 2/2
2018-12-14T09:14:49+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 22% 2/9
2018-12-14T09:15:11+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 33% 3/9
2018-12-14T09:15:44+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 44% 4/9
2018-12-14T09:23:03+01:00 P005419 167abc5b20e INFO RECOVERY progress of service: indexserver, abc-d-efg001:30040, volume: 2, 56% 5/9
<end>


Backint.log:
...
2018-12-14 09:22:58.563 backint started:
command: /usr/sap/XXX/SYS/global/hdb/opt/hdbbackint -f restore -p /usr/sap/XXX/SYS/global/hdb/opt/initSAP.utl -i /var/tmp/hdbbackint_XXX.6DYdDW -o /var/tmp/hdbbackint_XXX.d4OgZi -u XXX
pid: 12953
input:
#SOFTWAREID "backint 1.04" "HANA HDB server 2.00.024.06.1538035880"
#EBID VXP11544772783_abc-d-efg001 "/usr/sap/XXX/SYS/global/hdb/backint/DB_XXX_RECOVERY/log_backup_2_0_1993934784_2262273984"
<end>

The RecoveryLogReplayStarted & run until the indexserver crash occurred.


Indexserver trace:
...
5971]{-1}[-1/-1] 2018-12-14 09:33:00.532389 i Logger RecoveryHandlerImpl.cpp(00720) : Redo done up to position: 0x76d907c0 and time: 2018-12-14 08:01:00.307305+01:00 (29%)
[7886]{-1}[-1/-1] 2018-12-14 09:43:10.542330 i Logger RecoveryHandlerImpl.cpp(00720) : Redo done up to position: 0x76d907c0 and time: 2018-12-14 08:01:00.307305+01:00 (29%)
[7889]{-1}[-1/-1] 2018-12-14 09:44:38.371166 e Stream SynchronousPoolCopyHandler.cpp(00506) : updateTransferStats premature channel close: 0x00007f0ad8000000, state=Received, to transfer=134217728, transferred=389114, transferredChunks=0, channelClosed=1, signalTime=0, assignedChannel=ReadPipeChannel pipeAddress: /usr/sap/XXX/SYS/global/hdb/backint/DB_XXX_RECOVERY/log_backup_2_0_1993934784_2262273984
[7889]{-1}[-1/-1] 2018-12-14 09:44:38.371531 e Basis Crash.cpp(00756) : Crash at /data/jenkins_prod/workspace/85k332dv2a/s/DataAccess/DataRecovery/impl/RecoveryHandlerImpl.cpp:258

• SAP HANA DB 1.0
• SAP HANA DB 2.0

Bug, Symantec, Veritas, NetBackup , KBA , HAN-DB-BAC , SAP HANA Backup & Recovery , Problem