Hi,
we are using Xen3.0.1tha3 with kernel 2.6.12.6 on a debian maschine with
built in Broadvom tg3 NICs and the current debian testing maxdb. If
there is low traffic on the network everything runs fine and there no
obvious faults like dropped packets or other network errors. If we start
to generate higher traffic the database i.e. the whole virtual maschine
(domU) including the privileged domain (dom0) gets unreachable. We are
using the standard xen 3 bridging scripts and in the domU checksum
offloading(ethtool) is turned off.
Has someone see this problem as well. Any suggestions.
Thanks,
Artur
KERNELVERSION KERNEL 7.5.0 BUILD 031-000-DEB-001
INSTANCE_TYPE OLTP
MCOD YES
MAXLOGVOLUMES 2
MAXDATAVOLUMES 11
MAXBACKUPDEVS 2
LOG_MIRRORED NO
MAXCPU 4
MAXUSERTASKS 1400
MAXLOCKS 57280
CACHE_SIZE 80000
RUNDIRECTORY /var/lib/maxdb/data/wrk/PDB
OPMSG1 /dev/console
OPMSG2 /dev/null
LOG_SEGMENT_SIZE 200000
KERNELVERSION KERNEL 7.5.0 BUILD 031-000-DEB-001
INSTANCE_TYPE OLTP
MCOD YES
MAXLOGVOLUMES 2
MAXDATAVOLUMES 11
MAXBACKUPDEVS 2
LOG_MIRRORED NO
MAXCPU 4
MAXUSERTASKS 1400
MAXLOCKS 57280
CACHE_SIZE 80000
RUNDIRECTORY /var/lib/maxdb/data/wrk/PDB
OPMSG1 /dev/console
OPMSG2 /dev/null
LOG_SEGMENT_SIZE 200000
_UNICODE NO
DEFAULT_CODE ASCII
DATE_TIME_FORMAT INTERNAL
LOG_BACKUP_TO_PIPE NO
BACKUP_BLOCK_CNT 8
LOG_IO_QUEUE 50
_TRANS_RGNS 8
_TAB_RGNS 8
_ROW_RGNS 8
MAXSERVERTASKS 30
DEADLOCK_DETECTION 4
SESSION_TIMEOUT 900
REQUEST_TIMEOUT 5000
LRU_FOR_SCAN NO
CAT_CACHE_SUPPLY 44832
MP_RGN_LOOP 100
MAXRGN_REQUEST 3000
XP_DATA_CACHE_RGNS 0
_DATA_CACHE_RGNS 32
XP_CONVERTER_REGIONS 0
CONVERTER_REGIONS 8
XP_MAXPAGER 0
SEQUENCE_CACHE 1
VOLUMENO_BIT_COUNT 8
OPTIM_MAX_MERGE 500
OPTIM_INV_ONLY YES
JOIN_SEARCH_LEVEL 0
JOIN_MAXTAB_LEVEL4 16
JOIN_MAXTAB_LEVEL9 5
KERNELDIAGSIZE 800
_MAXEVENTS 200
TRACE_PAGES_TI 2
TRACE_PAGES_GC 0
TRACE_PAGES_LW 5
TRACE_PAGES_PG 3
TRACE_PAGES_US 10
TRACE_PAGES_UT 5
TRACE_PAGES_SV 5
TRACE_PAGES_EV 2
TRACE_PAGES_BUP 0
UTILITY_PROTSIZE 100
_MAX_MESSAGE_FILES 0
_EVENT_ALIVE_CYCLE 0
DIAG_HISTORY_NUM 2
DIAG_HISTORY_PATH /var/lib/maxdb/data/wrk/XPDB/DIAGHISTORY
_COMMENT
FORMATTING_MODE PARALLEL
FORMAT_DATAVOLUME YES
OFFICIAL_NODE
LOAD_BALANCING_CHK 0
HS_STORAGE_DLL libhsscopy
HS_NODE_?
HS_DELAY_TIME_?
HS_SYNC_INTERVAL 50
JOIN_OPERATOR_IMPLEMENTATION YES
JOIN_TABLEBUFFER 128
SHAREDSQL NO
SHAREDSQL_EXPECTEDSTATEMENTCOUNT 1500
SHAREDSQL_COMMANDCACHESIZE 32768
MEMORY_ALLOCATION_LIMIT 0
MIN_RETENTION_TIME 60
MAX_RETENTION_TIME 480
MAX_SINGLE_HASHTABLE_SIZE 512
MAX_HASHTABLE_MEMORY 5120
HASHED_RESULTSET YES
HASHED_RESULTSET_CACHESIZE 262144
AUTO_RECREATE_BAD_INDEXES NO
LOCAL_REDO_LOG_BUFFER_SIZE 0
--- Begin Message ---
Hi,
We are trying to connect more than two JBoss-instances with one instance
of maxdb(on Linux). We first ran into problems when trying to connect
the second JBoss, but were able to solve them by increasing the value of
MAX_USER_TASKS to 380. When we tried to connect the 3rd JBoss the
database says "could not create communication semaphore" and rejects all
further connection attemps by JBoss's JDBC driver. We then increased the
MAX_USER_TASKS value by tweaking the maximum shared memory value with
"sysctl -w kernel.shmmax=134217738" to 1400, but that didn't help. MaxDB
still crashes at about 150 sessions opened by the JBoss-servers.
So any suggestions how to solve this (see at the end for details about
our current parameters/setup)?
Thanks in advance,
Artur
We are currently using the follwoing setup:
-one instance of maxdb running with 2 GB memory and 2x3.2 GHz Xeon CPus
with Hyperthreading on Linux 2.6.15.1
-all JBoss-Servers on XEN 2 Virtual Machines with 830 MB of memory in
the same subnet
-http-session loadbalancing with apache-ssl/mod-jk
KERNELVERSION KERNEL 7.5.0 BUILD 031-000-DEB-001
INSTANCE_TYPE OLTP
MCOD YES
MAXLOGVOLUMES 2
MAXDATAVOLUMES 11
MAXBACKUPDEVS 2
LOG_MIRRORED NO
MAXCPU 4
MAXUSERTASKS 1400
MAXLOCKS 57280
CACHE_SIZE 80000
RUNDIRECTORY /var/lib/maxdb/data/wrk/PDB
OPMSG1 /dev/console
OPMSG2 /dev/null
LOG_SEGMENT_SIZE 200000
KERNELVERSION KERNEL 7.5.0 BUILD 031-000-DEB-001
INSTANCE_TYPE OLTP
MCOD YES
MAXLOGVOLUMES 2
MAXDATAVOLUMES 11
MAXBACKUPDEVS 2
LOG_MIRRORED NO
MAXCPU 4
MAXUSERTASKS 1400
MAXLOCKS 57280
CACHE_SIZE 80000
RUNDIRECTORY /var/lib/maxdb/data/wrk/PDB
OPMSG1 /dev/console
OPMSG2 /dev/null
LOG_SEGMENT_SIZE 200000
_UNICODE NO
DEFAULT_CODE ASCII
DATE_TIME_FORMAT INTERNAL
LOG_BACKUP_TO_PIPE NO
BACKUP_BLOCK_CNT 8
LOG_IO_QUEUE 50
_TRANS_RGNS 8
_TAB_RGNS 8
_ROW_RGNS 8
MAXSERVERTASKS 30
DEADLOCK_DETECTION 4
SESSION_TIMEOUT 900
REQUEST_TIMEOUT 5000
LRU_FOR_SCAN NO
CAT_CACHE_SUPPLY 44832
MP_RGN_LOOP 100
MAXRGN_REQUEST 3000
XP_DATA_CACHE_RGNS 0
_DATA_CACHE_RGNS 32
XP_CONVERTER_REGIONS 0
CONVERTER_REGIONS 8
XP_MAXPAGER 0
SEQUENCE_CACHE 1
VOLUMENO_BIT_COUNT 8
OPTIM_MAX_MERGE 500
OPTIM_INV_ONLY YES
JOIN_SEARCH_LEVEL 0
JOIN_MAXTAB_LEVEL4 16
JOIN_MAXTAB_LEVEL9 5
KERNELDIAGSIZE 800
_MAXEVENTS 200
TRACE_PAGES_TI 2
TRACE_PAGES_GC 0
TRACE_PAGES_LW 5
TRACE_PAGES_PG 3
TRACE_PAGES_US 10
TRACE_PAGES_UT 5
TRACE_PAGES_SV 5
TRACE_PAGES_EV 2
TRACE_PAGES_BUP 0
UTILITY_PROTSIZE 100
_MAX_MESSAGE_FILES 0
_EVENT_ALIVE_CYCLE 0
DIAG_HISTORY_NUM 2
DIAG_HISTORY_PATH /var/lib/maxdb/data/wrk/XPDB/DIAGHISTORY
_COMMENT
FORMATTING_MODE PARALLEL
FORMAT_DATAVOLUME YES
OFFICIAL_NODE
LOAD_BALANCING_CHK 0
HS_STORAGE_DLL libhsscopy
HS_NODE_?
HS_DELAY_TIME_?
HS_SYNC_INTERVAL 50
JOIN_OPERATOR_IMPLEMENTATION YES
JOIN_TABLEBUFFER 128
SHAREDSQL NO
SHAREDSQL_EXPECTEDSTATEMENTCOUNT 1500
SHAREDSQL_COMMANDCACHESIZE 32768
MEMORY_ALLOCATION_LIMIT 0
MIN_RETENTION_TIME 60
MAX_RETENTION_TIME 480
MAX_SINGLE_HASHTABLE_SIZE 512
MAX_HASHTABLE_MEMORY 5120
HASHED_RESULTSET YES
HASHED_RESULTSET_CACHESIZE 262144
AUTO_RECREATE_BAD_INDEXES NO
LOCAL_REDO_LOG_BUFFER_SIZE 0
--- End Message ---
--
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
