RAC interconnect and crossover
Let assume that we have only 2 nodes in oracle cluster.
What is better for interconnect:
Crossover or network with switch and why?
Thank you.
ps. we have dispute with 2nd DBA.
Hi Alex, all,
sorry, but I find this discussion dangerous, since missleading. Please, let me make it very clear:
If you plan to set up a production system, Oracle will not support crossover cables!
Therefore, regardless of whether or not a crossover cable is technically possible, it has no advantage over a switch. For test systems you may want to use anything that you want. However, please, note that a test system with a crossover cable cannot be a test environment for interconnect failure tests.
The reason is that Oracle has found that a lot of the interconnect failure tests will perform totally differently when using a switch - the network guys on this forum may want to elaborate why. In addition, Oracle has seen issues with crossover cables even under normal circumstances (no physical failure). While those were mostly based on auto speed negotiation of the NICs on both sides when using a crossover cable or other OS settings that need to be optimized, the behaviour was more predictable and stable using a switch.
Concluding, you would need to use a switch on your interconnect.
Thanks,
Markus
Similar Messages
-
RAC, Interconnect and CloudControl (a nice to have problem :-) )
Hi,
someone has changed the networkmask for one interconnect interface (eth3). eth3 is up an running, the interconnect ip-address move to the second interconnect interface here eth1. Everything is working fine, and after network is changed back the eth3 interface get his interconnect ip-address and all is fine. But there was no incident in CloudControl no warning that something is going wrong. A user defined metric on GV_$CLUSTER_INTERCONNECTS is no solution, the view shows old values. cluyfy is useful in this case.
best regards
ThomasHi,
I think two options
cluvfy and oifcfg to check the network configuration
regards -
Oracle RAC Interconnect, PowerVM VLANs, and the Limit of 20
Hello,
Our company has a requirement to build a multitude of Oracle RAC clusters on AIX using Power VM on 770s and 795 hardware.
We presently have 802.1q trunking configured on our Virtual I/O Servers, and have currently consumed 12 of 20 allowed VLANs for a virtual ethernet adapter. We have read the Oracle RAC FAQ on Oracle Metalink and it seems to otherwise discourage the use of sharing these interconnect VLANs between different clusters. This puts us in a scalability bind; IBM limits VLANs to 20 and Oracle says there is a one-to-one relationship between VLANs and subnets and RAC clusters. We must assume we have a fixed number of network interfaces available and that we absolutely have to leverage virtualized network hardware in order to build these environments. "add more network adapters to VIO" isn't an acceptable solution for us.
Does anyone know if Oracle can afford any flexibility which would allow us to host multiple Oracle RAC interconnects on the same 802.1q trunk VLAN? We will independently guarantee the bandwidth, latency, and redundancy requirements are met for proper Oracle RAC performance, however we don't want a design "flaw" to cause us supportability issues in the future.
We'd like it very much if we could have a bunch of two-node clusters all sharing the same private interconnect. For example:
Cluster 1, node 1: 192.168.16.2 / 255.255.255.0 / VLAN 16
Cluster 1, node 2: 192.168.16.3 / 255.255.255.0 / VLAN 16
Cluster 2, node 1: 192.168.16.4 / 255.255.255.0 / VLAN 16
Cluster 2, node 2: 192.168.16.5 / 255.255.255.0 / VLAN 16
Cluster 3, node 1: 192.168.16.6 / 255.255.255.0 / VLAN 16
Cluster 3, node 2: 192.168.16.7 / 255.255.255.0 / VLAN 16
Cluster 4, node 1: 192.168.16.8 / 255.255.255.0 / VLAN 16
Cluster 4, node 2: 192.168.16.9 / 255.255.255.0 / VLAN 16
etc.
Whereas the concern is that Oracle Corp will only support us if we do this:
Cluster 1, node 1: 192.168.16.2 / 255.255.255.0 / VLAN 16
Cluster 1, node 2: 192.168.16.3 / 255.255.255.0 / VLAN 16
Cluster 2, node 1: 192.168.17.2 / 255.255.255.0 / VLAN 17
Cluster 2, node 2: 192.168.17.3 / 255.255.255.0 / VLAN 17
Cluster 3, node 1: 192.168.18.2 / 255.255.255.0 / VLAN 18
Cluster 3, node 2: 192.168.18.3 / 255.255.255.0 / VLAN 18
Cluster 4, node 1: 192.168.19.2 / 255.255.255.0 / VLAN 19
Cluster 4, node 2: 192.168.19.3 / 255.255.255.0 / VLAN 19
Which eats one VLAN per RAC cluster.Thank you for your answer!!
I think I roughly understand the argument behind a 2-node RAC and a 3-node or greater RAC. We, unfortunately, were provided with two physical pieces of hardware to virtualize to support production (and two more to support non-production) and as a result we really have no place to host a third RAC node without placing it within the same "failure domain" (I hate that term) as one of the other nodes.
My role is primarily as a system engineer, and, generally speaking, our main goals are eliminating single points of failure. We may be misusing 2-node RACs to eliminate single points of failure since it seems to violate the real intentions behind RAC, which is used more appropriately to scale wide to many nodes. Unfortunately, we've scaled out to only two nodes, and opted to scale these two nodes up, making them huge with many CPUs and lots of memory.
Other options, notably the active-passive failover cluster we have in HACMP or PowerHA on the AIX / IBM Power platform is unattractive as the standby node drives no resources yet must consume CPU and memory resources so that it is prepared for a failover of the primary node. We use HACMP / PowerHA with Oracle and it works nice, however Oracle RAC, even in a two-node configuration, drives load on both nodes unlike with an active-passive clustering technology.
All that aside, I am posing the question to both IBM, our Oracle DBAs (whom will ask Oracle Support). Typically the answers we get vary widely depending on the experience and skill level of the support personnel we get on both the Oracle and IBM sides... so on a suggestion from a colleague (Hi Kevin!) I posted here. I'm concerned that the answer from Oracle Support will unthinkingly be "you can't do that, my script says to tell you the absolute most rigid interpretation of the support document" while all the time the same document talks of the use of NFS and/or iSCSI storage eye roll
We have a massive deployment of Oracle EBS and honestly the interconnect doesn't even touch 100mbit speeds even though the configuration has been checked multiple times by Oracle and IBM and with the knowledge that Oracle EBS is supposed to heavily leverage RAC. I haven't met a single person who doesn't look at our environment and suggest jumbo frames. It's a joke at this point... comments like "OMG YOU DON'T HAVE JUMBO FRAMES" and/or "OMG YOU'RE NOT USING INFINIBAND WHATTA NOOB" are commonplace when new DBAs are hired. I maintain that the utilization numbers don't support this.
I can tell you that we have 8Gb fiber channel storage and 10Gb network connectivity. I would probably assume that there were a bottleneck in the storage infrastructure first. But alas, I digress.
Mainly I'm looking for a real-world answer to this question. Aside from violating every last recommendation and making oracle support folk gently weep at the suggestion, are there any issues with sharing interconnects between RAC environments that will prevent it's functionality and/or reduce it's stability?
We have rapid spanning tree configured, as far as I know, and our network folks have tuned the timers razor thin. We have Nexus 5k and Nexus 7k network infrastructure. The typical issues you'd fine with standard spanning tree really don't affect us because our network people are just that damn good. -
10g RAC HBA and Interconnect Adapters
Hi,
I am setting up an Oracle 10g RAC environment on two Sun 220R servers. The shared storage system will be a an EMC Clarion SAN (CX600/CX700 model drives). I have two questions below.
'Supported Interconnects' and 'Public Networks' certified configuration adapters include '... and SunSwift(tm) adapters'. Does this mean any/all SunSwift adapter? Specifically I am looking at the 'Quad GigaSwift ethernet UTP adapter'.
I have not seen any configuration include certified HBA adapters. Does this imply so long as Sun is happy with it, Oracle 10g RAC will be happy with it as well?
Many thanks in advance,
Jack.Sorry for the late reply, but the interface to pick is the clprivnet0. This load-balances over the available private interconnects under the covers and so does not represent a single point of failure.
Tim
--- -
802.3ad (mode=4) bonding for RAC interconnects
Is anyone using 802.3ad (mode=4) bonding for their RAC interconnects? We have five Dell R710 RAC nodes and we're trying to use the four onboard Broadcom NetXtreme II NICs in a 802.3ad bond with src-dst-mac load balancing. Since we have the hardware to pull this off we thought we'd give it a try and achieve some extra bandwith for the interconnect rather than deploying the traditional acitve/standby interconnect using just two of the NICs. Has anyone tried this config and what was the outcome? Thanks.
I don't but may be the documents might help ?
http://www.iop.org/EJ/article/1742-6596/119/4/042015/jpconf8_119_042015.pdf?request-id=bcddc94d-7727-4a8a-8201-4d1b837a1eac
http://www.oracleracsig.org/pls/apex/Z?p_url=RAC_SIG.download_my_file?p_file=1002938&p_id=1002938&p_cat=documents&p_user=nobody&p_company=994323795175833
http://www.oracle.com/technology/global/cn/events/download/ccb/10g_rac_bp_en.pdf
Edited by: Hub on Nov 18, 2009 10:10 AM -
Hi guys,
I've this doubt since I'm using Oracle RAC... Oracle recommends the use of a dedicated swith to the RAC servers communication. We are talking about using a dedicated switch to the heartbeat mechanism or to allowing one RAC node to read data blocks from another RAC node's cache via the Interconnect?
Thanks.
A.krishan Jaglan wrote:
Hi User8897201,
as far as i know, Interconnect and heartbeat goes over same link( there is no way to separate them out). If you have Big switch and/Or don't want a physical switch just of RAC , you can ask your Network Administrator to create VLAN for (Interconnect ) - Its same as dedicated switch as RAC traffic will no interrupted by flooded network by anything else.
Key is Network must be dedicated for Interconnect( Non Shared Network), how you achieve doesn't mattter ( either dedicated switch or VLAN just for Interconnect).Again, I have seen a switch that uses a VLAN get saturated to the point that one or more nodes from a cluster rebooted. A switch - even a director-class switch only has so much bandwidth - which if not dedicated to cluster interconnect can still get overwhelmed causing node evictions. Been there, done that, switched to dedicated and then got the T-Shirt.
>
Hope it helps you.
Thanks
krishan Jaglan -
Hi,
We are facing RAC Interconnect performance problems.
Oracle Version: Oracle 9i RAC (9.2.0.7)
Operating system: SunOS 5.8
SQL> SELECT b1.inst_id, b2.value "RECEIVED",
b1.value "RECEIVE TIME",
((b1.value / b2.value) * 10) "AVG RECEIVE TIME (ms)"
FROM gv$sysstat b1, gv$sysstat b2
WHERE b1.name = 'global cache cr block receive time'
AND b2.name = 'global cache cr blocks received'
AND b1.inst_id = b2.inst_id;
INST_ID RECEIVED RECEIVE TIME AVG RECEIVE TIME (ms)
1 323849 172359 5.32220263
2 675806 94537 1.39887778
After database restart average time increases for Instance 1 and instance 2 remains similar.
Application performance degrades, restart database solves the issue. This is critical application and can not have frequent downtimes for restart.
What specific points should I check to find out to improve interconnect performance?
Thanks
Dilip Patel.Hi,
Configurations:
Node: 1
Hardware Model: Sun-Fire-V890
OS: SunOS 5.8
Release: Generic_117350-53
CPU: 16 sparcv9 cpu(s) running at 1200 MHz
Memory: 40.0GB
Node: 2
Hardware Model: Sun-Fire-V890
OS: SunOS 5.8
Release: Generic_117350-53
CPU: 16 sparcv9 cpu(s) running at 1200 MHz
Memory: 40.0GB
CPU Utilization on Node 1 is never exceeded 40%.
CPU Utilization on Node 2 is between 20% to 30%.
Application load is more Node 1 compared to Node 2.
I can observer wait event "global cache cr request" in top 5 wait events on most of the statspack report. Application faces degrade performacne after few days of restart database. No major changes done on application recently.
Statapack report for Node 1:
DB Name DB Id Instance Inst Num Release Cluster Host
XXXX 2753907139 xxxx1 1 9.2.0.7.0 YES xxxxx
Snap Id Snap Time Sessions Curs/Sess Comment
Begin Snap: 61688 17-Feb-09 09:10:06 253 299.4
End Snap: 61698 17-Feb-09 10:10:06 285 271.6
Elapsed: 60.00 (mins)
Cache Sizes (end)
~~~~~~~~~~~~~~~~~
Buffer Cache: 2,048M Std Block Size: 8K
Shared Pool Size: 384M Log Buffer: 2,048K
Load Profile
~~~~~~~~~~~~ Per Second Per Transaction
Redo size: 102,034.92 4,824.60
Logical reads: 60,920.35 2,880.55
Block changes: 986.07 46.63
Physical reads: 1,981.12 93.67
Physical writes: 28.30 1.34
User calls: 2,651.63 125.38
Parses: 500.89 23.68
Hard parses: 21.44 1.01
Sorts: 66.91 3.16
Logons: 3.69 0.17
Executes: 553.34 26.16
Transactions: 21.15
% Blocks changed per Read: 1.62 Recursive Call %: 22.21
Rollback per transaction %: 2.90 Rows per Sort: 7.44
Instance Efficiency Percentages (Target 100%)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Buffer Nowait %: 99.99 Redo NoWait %: 100.00
Buffer Hit %: 96.75 In-memory Sort %: 100.00
Library Hit %: 98.30 Soft Parse %: 95.72
Execute to Parse %: 9.48 Latch Hit %: 99.37
Parse CPU to Parse Elapsd %: 90.03 % Non-Parse CPU: 92.97
Shared Pool Statistics Begin End
Memory Usage %: 94.23 94.93
% SQL with executions>1: 74.96 74.66
% Memory for SQL w/exec>1: 82.93 72.26
Top 5 Timed Events
~~~~~~~~~~~~~~~~~~ % Total
Event Waits Time (s) Ela Time
db file sequential read 1,080,532 13,191 40.94
CPU time 10,183 31.60
db file scattered read 456,075 3,977 12.34
wait for unread message on broadcast channel 4,195 2,770 8.60
global cache cr request 1,633,056 873 2.71
Cluster Statistics for DB: EPIP Instance: epip1 Snaps: 61688 -61698
Global Cache Service - Workload Characteristics
Ave global cache get time (ms): 0.8
Ave global cache convert time (ms): 1.1
Ave build time for CR block (ms): 0.1
Ave flush time for CR block (ms): 0.2
Ave send time for CR block (ms): 0.3
Ave time to process CR block request (ms): 0.6
Ave receive time for CR block (ms): 4.4
Ave pin time for current block (ms): 0.2
Ave flush time for current block (ms): 0.0
Ave send time for current block (ms): 0.3
Ave time to process current block request (ms): 0.5
Ave receive time for current block (ms): 2.6
Global cache hit ratio: 3.9
Ratio of current block defers: 0.0
% of messages sent for buffer gets: 3.7
% of remote buffer gets: 0.3
Ratio of I/O for coherence: 1.1
Ratio of local vs remote work: 10.9
Ratio of fusion vs physical writes: 0.0
Global Enqueue Service Statistics
Ave global lock get time (ms): 0.1
Ave global lock convert time (ms): 0.0
Ratio of global lock gets vs global lock releases: 1.0
GCS and GES Messaging statistics
Ave message sent queue time (ms): 0.4
Ave message sent queue time on ksxp (ms): 1.8
Ave message received queue time (ms): 0.2
Ave GCS message process time (ms): 0.1
Ave GES message process time (ms): 0.0
% of direct sent messages: 8.0
% of indirect sent messages: 49.4
% of flow controlled messages: 42.6
GES Statistics for DB: EPIP Instance: epip1 Snaps: 61688 -61698
Statistic Total per Second per Trans
dynamically allocated gcs resourc 0 0.0 0.0
dynamically allocated gcs shadows 0 0.0 0.0
flow control messages received 0 0.0 0.0
flow control messages sent 0 0.0 0.0
gcs ast xid 0 0.0 0.0
gcs blocked converts 2,830 0.8 0.0
gcs blocked cr converts 7,677 2.1 0.1
gcs compatible basts 5 0.0 0.0
gcs compatible cr basts (global) 142 0.0 0.0
gcs compatible cr basts (local) 142,678 39.6 1.9
gcs cr basts to PIs 0 0.0 0.0
gcs cr serve without current lock 0 0.0 0.0
gcs error msgs 0 0.0 0.0
gcs flush pi msgs 798 0.2 0.0
gcs forward cr to pinged instance 0 0.0 0.0
gcs immediate (compatible) conver 9,296 2.6 0.1
gcs immediate (null) converts 52,460 14.6 0.7
gcs immediate cr (compatible) con 752,507 209.0 9.9
gcs immediate cr (null) converts 4,047,959 1,124.4 53.2
gcs msgs process time(ms) 153,618 42.7 2.0
gcs msgs received 2,287,640 635.5 30.0
gcs out-of-order msgs 0 0.0 0.0
gcs pings refused 70,099 19.5 0.9
gcs queued converts 0 0.0 0.0
gcs recovery claim msgs 0 0.0 0.0
gcs refuse xid 1 0.0 0.0
gcs retry convert request 0 0.0 0.0
gcs side channel msgs actual 40,400 11.2 0.5
gcs side channel msgs logical 4,039,700 1,122.1 53.1
gcs write notification msgs 46 0.0 0.0
gcs write request msgs 972 0.3 0.0
gcs writes refused 4 0.0 0.0
ges msgs process time(ms) 2,713 0.8 0.0
ges msgs received 73,687 20.5 1.0
global posts dropped 0 0.0 0.0
global posts queue time 0 0.0 0.0
global posts queued 0 0.0 0.0
global posts requested 0 0.0 0.0
global posts sent 0 0.0 0.0
implicit batch messages received 288,801 80.2 3.8
implicit batch messages sent 622,610 172.9 8.2
lmd msg send time(ms) 2,148 0.6 0.0
lms(s) msg send time(ms) 1 0.0 0.0
messages flow controlled 3,473,393 964.8 45.6
messages received actual 765,292 212.6 10.1
messages received logical 2,360,972 655.8 31.0
messages sent directly 654,760 181.9 8.6
messages sent indirectly 4,027,924 1,118.9 52.9
msgs causing lmd to send msgs 33,481 9.3 0.4
msgs causing lms(s) to send msgs 13,220 3.7 0.2
msgs received queue time (ms) 379,304 105.4 5.0
msgs received queued 2,359,723 655.5 31.0
msgs sent queue time (ms) 1,514,305 420.6 19.9
msgs sent queue time on ksxp (ms) 4,349,174 1,208.1 57.1
msgs sent queued 4,032,426 1,120.1 53.0
msgs sent queued on ksxp 2,415,381 670.9 31.7
GES Statistics for DB: EPIP Instance: epip1 Snaps: 61688 -61698
Statistic Total per Second per Trans
process batch messages received 278,174 77.3 3.7
process batch messages sent 913,611 253.8 12.0
Wait Events for DB: EPIP Instance: epip1 Snaps: 61688 -61698
-> s - second
-> cs - centisecond - 100th of a second
-> ms - millisecond - 1000th of a second
-> us - microsecond - 1000000th of a second
-> ordered by wait time desc, waits desc (idle events last)
Avg
Total Wait wait Waits
Event Waits Timeouts Time (s) (ms) /txn
db file sequential read 1,080,532 0 13,191 12 14.2
db file scattered read 456,075 0 3,977 9 6.0
wait for unread message on b 4,195 1,838 2,770 660 0.1
global cache cr request 1,633,056 8,417 873 1 21.4
db file parallel write 8,243 0 260 32 0.1
buffer busy waits 16,811 0 168 10 0.2
log file parallel write 187,783 0 158 1 2.5
log file sync 75,143 0 147 2 1.0
buffer busy global CR 9,713 0 102 10 0.1
global cache open x 31,157 1,230 50 2 0.4
enqueue 58,261 14 45 1 0.8
latch free 33,398 7,610 44 1 0.4
direct path read (lob) 9,925 0 36 4 0.1
library cache pin 8,777 1 34 4 0.1
SQL*Net break/reset to clien 82,982 0 32 0 1.1
log file sequential read 409 0 31 75 0.0
log switch/archive 3 3 29 9770 0.0
SQL*Net more data to client 201,538 0 16 0 2.6
global cache open s 8,585 342 14 2 0.1
global cache s to x 11,098 148 11 1 0.1
control file sequential read 6,845 0 8 1 0.1
db file parallel read 1,569 0 7 4 0.0
log file switch completion 35 0 7 194 0.0
row cache lock 15,780 0 6 0 0.2
process startup 69 0 6 82 0.0
global cache null to x 1,759 48 6 3 0.0
direct path write (lob) 685 0 5 7 0.0
DFS lock handle 8,713 0 3 0 0.1
control file parallel write 1,350 0 2 2 0.0
wait for master scn 1,194 0 1 1 0.0
CGS wait for IPC msg 30,830 30,715 1 0 0.4
global cache busy 14 1 1 75 0.0
ksxr poll remote instances 30,997 12,692 1 0 0.4
direct path read 752 0 0 1 0.0
switch logfile command 3 0 0 148 0.0
log file single write 24 0 0 13 0.0
library cache lock 668 0 0 0 0.0
KJC: Wait for msg sends to c 1,161 0 0 0 0.0
buffer busy global cache 26 0 0 6 0.0
IPC send completion sync 261 260 0 0 0.0
PX Deq: reap credit 3,477 3,440 0 0 0.0
LGWR wait for redo copy 1,751 0 0 0 0.0
async disk IO 1,059 0 0 0 0.0
direct path write 298 0 0 0 0.0
slave TJ process wait 1 1 0 18 0.0
PX Deq: Execute Reply 3 1 0 3 0.0
PX Deq: Join ACK 8 4 0 1 0.0
global cache null to s 8 0 0 1 0.0
ges inquiry response 16 0 0 0 0.0
Wait Events for DB: EPIP Instance: epip1 Snaps: 61688 -61698
-> s - second
-> cs - centisecond - 100th of a second
-> ms - millisecond - 1000th of a second
-> us - microsecond - 1000000th of a second
-> ordered by wait time desc, waits desc (idle events last)
Avg
Total Wait wait Waits
Event Waits Timeouts Time (s) (ms) /txn
PX Deq: Parse Reply 6 2 0 1 0.0
PX Deq Credit: send blkd 2 1 0 0 0.0
PX Deq: Signal ACK 3 1 0 0 0.0
library cache load lock 1 0 0 0 0.0
buffer deadlock 6 6 0 0 0.0
lock escalate retry 4 4 0 0 0.0
SQL*Net message from client 9,470,867 0 643,285 68 124.4
queue messages 42,829 41,144 42,888 1001 0.6
wakeup time manager 601 600 16,751 27872 0.0
gcs remote message 795,414 120,163 13,606 17 10.4
jobq slave wait 2,546 2,462 7,375 2897 0.0
PX Idle Wait 2,895 2,841 7,021 2425 0.0
virtual circuit status 120 120 3,513 29273 0.0
ges remote message 142,306 69,912 3,504 25 1.9
SQL*Net more data from clien 206,559 0 19 0 2.7
SQL*Net message to client 9,470,903 0 14 0 124.4
PX Deq: Execution Msg 313 103 2 7 0.0
Background Wait Events for DB: EPIP Instance: epip1 Snaps: 61688 -61698
-> ordered by wait time desc, waits desc (idle events last)
Avg
Total Wait wait Waits
Event Waits Timeouts Time (s) (ms) /txn
db file parallel write 8,243 0 260 32 0.1
log file parallel write 187,797 0 158 1 2.5
log file sequential read 316 0 22 70 0.0
enqueue 56,204 0 15 0 0.7
control file sequential read 5,694 0 6 1 0.1
DFS lock handle 8,682 0 3 0 0.1
db file sequential read 276 0 2 8 0.0
control file parallel write 1,334 0 2 2 0.0
wait for master scn 1,194 0 1 1 0.0
CGS wait for IPC msg 30,830 30,714 1 0 0.4
ksxr poll remote instances 30,972 12,681 1 0 0.4
latch free 356 54 1 2 0.0
direct path read 752 0 0 1 0.0
log file single write 24 0 0 13 0.0
LGWR wait for redo copy 1,751 0 0 0 0.0
async disk IO 812 0 0 0 0.0
global cache cr request 69 0 0 1 0.0
row cache lock 45 0 0 1 0.0
direct path write 298 0 0 0 0.0
library cache pin 29 0 0 1 0.0
rdbms ipc reply 29 0 0 0 0.0
buffer busy waits 10 0 0 0 0.0
library cache lock 2 0 0 0 0.0
global cache open x 2 0 0 0 0.0
rdbms ipc message 179,764 36,258 29,215 163 2.4
gcs remote message 795,409 120,169 13,605 17 10.4
pmon timer 1,388 1,388 3,508 2527 0.0
ges remote message 142,295 69,912 3,504 25 1.9
smon timer 414 0 3,463 8366 0.0
------------------------------------------------------------- -
RAC Interconnect Transfer rate vs NIC's Bandwidth
Hi Guru,
I need some clarification for RAC interconnect terminology between "private interconnect transfer rate" and "NIC bandwidth".
We have 11gR2 RAC with multiple databases.
So we need to find out what the current resource status is.
We have two physical NICs each node. And 8G is for public and 2G is for private (interconnect).
Technically, we have 4G for Private network bandwidth.
If I look at the "Private Interconnect Transfer rate" though OEM or IPTraf (linux tool), it is showing 20 ~30 MB/Sec.
There is no any issue at all at this moment.
Please correct me if I am wrong.
The transfer rate will be fine till 500M or 1G/Sec. Because the current NIC's capacity is 4G. Does it make sense ?
I'm sure there are multiple things to consider,but I'm kind of stumped on the whole transfer rate vs bandwidth. Is there any way to calculate what a typical transfer would be....
OR How do I say our interconnect are good enough ....based on the transfer rate ?
Another question is ....
In our case, how do I set up the warning threshold and Critical threshold for "Private Interconnect Transer rate" in OEM ?
Any comments will be appreciated.
Please advise.Interconnect performance sways more to latency than bandwidth IMO. In simplistic terms, memory is shared across the Interconnect. What is important for accessing memory? The size of the pipe? Or the speed of the pipe?
A very fast small pipe will typically perform significantly better than a large and slower pipe.
Even the size of the pipe is not that straight forward. Standard IP MTU size is 1500. You can run jumbo and super-jumbo frame MTU sizes on the Interconnect - where for example a MTU size of 65K is significantly larger than a 1500 byte MTU. Which means significantly more data can be transferred over the Interconnect at a much reduced overhead.
Personally, I would not consider Ethernet (GigE included) for the Interconnect. Infiniband is faster, more scalable, and offers an actual growth path to 128Gb/s and higher.
Oracle also uses Infiniband (QDR/40Gb) for their Exadata Database Machine product's Interconnect. Infiniband also enables one to run Oracle Interconnect over RDS instead of UDP. I've seen Oracle reports to the OFED committee saying that using RDS in comparison with UDP, reduced CPU utilisation by 50% and decreased latency by 50%.
I also do not see the logic of having a faster public network and a slower Interconnect.
IMO there are 2 very fundamental components in RAC that determines what is the speed and performance achievable with that RAC - the speed, performance and scalability of the I/O fabric layer and for the Interconnect layer.
And Exadata btw uses Infiniband for both these critical layers. Not fibre. Not GigE. -
We have set up a RAC DB and we are trying to understand the effect of interconnect failure. Our understanding is that when interconnect fails, the RAC will use the information in voting disk to decide which node will survive and which node will be shut down. We want to make sure that is what is going to happen. Interconnect failure will not result in both node being shutdown, correct?
Please point us to the right oracle doc about this isssue, if one exists. We were not able to find firm answer on the docs we have read so far.
Thanks.Is it true that when a interconnect failure occurs in a two node cluster, lowest node number rule? Well, I am not sure this could be true or not. Couple of times its been observed in our two node RAC Databases where each time one node evicted due to couple reasons (one of the reason was the time synchronization between the nodes) and surprisingly another time a node evicted without any valid reason too.
Jaffar -
In the client place the architecture team wants to implement a a node RAC cluster on Sun Solaris on Oracle 10g. But in order to minimize the interconnect traffic they want applications to connect only to one node and the other node will provide fail over capability.
Though this can be achieved with less complexity thru physical standby, theoretically if in a 2 node RAC cluster if we just use one node will it eliminate interconnect traffic?
I guess that there will not be any requests from the second node for blocks held by the first node so no transfer of blocks over interconnect will take place. But still the second node has to know about the blocks held by the first node in order to recover in case the first node crashes(cache fusion?). Right?
TIA
RadKrishjust my 0.02 cents on this .. Why would you want to invest in a High Availability Solution ( RAC ) and attempt to achieve Disaster Recovery ( Failover ) through it ..
Why would you want to minimize the Interconnect traffic in the first place .. A decent Switch plus Gigabit Ethernet / Cat5 cables is all you need for the Interconnect and its known to work well with most types of RAC Setup's ..
With Dynamic Resource Remastering , mastership of blocks across the active nodes get distributed based on access patterns .. so with your kind of setup , almost all the blocks would be mastered on Node:1 after a period of sustained activity .. Further Instance recovery happens using information from the Redo Log Thread of the crashed Node so it should not affect Interconnect traffic ..
Vishwa -
Dedicated switches needed for RAC interconnect or not?
Currently working on an Extended RAC cluster design implementation, I asked the network engineer for dedicated switches for the RAC interconnects.
Here is a little background:
There are 28 RAC clusters over 2X13 physical RAC nodes with separate Oracle_Home for each instance with atleast 2+ instances on each RAC node. So 13 RAC nodes will be in each site(Data-Center). This is basically an Extended RAC solution for SAP databases on RHEL 6 using ASM and Clusterware for Oracle 11gR2. The RAC nodes are Blades in a c7000 enclosure (in each site). The distance between the sites is 55+ kms.
Oracle recommends to have Infiniband(20GBps) as the network backbone, but here DWDM will be used with 2X10 Gbps (each at 10 GBps) links for the RAC interconnect between the sites. There will be separate 2x1GBps redundant link for the Production network and 2x2 GBps FC(Fiber-Channel) redundant links for the SAN/Storage(ASM traffic will go here) network. There will be switches for the Public-production network and the SAN network each.
Oracle recommends dedicated switches(which will give acceptable latency/bandwith) with switch redundancy to route the dedicated/non-routable VLANs for the RAC interconnect (private/heartbeat/global cache transfer) network. Since the DWDM interlinks is 2x10Gbps - do I still need the dedicated switches?
If yes, then how many?
Your inputs will be greatly appreciated.. and help me take a decision.
Many Thanks in advance..
AbhijitAbsolutely agree.. the chances of overload in a HA(RAC) solution and ultmate RAC node eviction are very high(with very high latency) and for exactly this reason I even suggested inexpensive switches to route the VLANs for the RAC interconnect through these switches. The ASM traffic will get routed through the 2x2GB FC links through SAN-Directors (1 in each site).
Suggested the network folks to use Up-links from the c7000 enclosure and route the RAC VLAN through these inexpensive switches for the interconnect traffic. We have another challenge here: HP has certified using VirtualConnect/Flex-Fabric architecture for Blades in c7000 to allocate VLANs for RAC interconnect. But this is only for one site, and does not span Production/DR sites separated over a distance.
Btw, do you have any standard switch model to select from.. and how many to go for a RAC configuration of 13 Extended RAC clusters with each cluster hosting 2+ RAC instances to host total of 28 SAP instances.
Many Thanks again!
Abhijit -
Oracle 11gR2 RAC VM and SCAN and DNS and /etc/hosts (two) setup questions
Hi,
I am looking forward to setting up two Oracle 11gR2 RAC instances
on my Oracle VM test machine.
I plan on using the Oracle 11gR2 RAC VM template.
I want the final Oracle 11gR2 RAC instances to have SCAN that uses DNS.
The DNS will be pre-installed in the JeOS.
My first simple question about the setup is the following.
In my DNS name file, for example,
/var/named/chroot/var/named/milkyway.univ.db
do I need to provide the racnode1 and racnode2 information,
for example,
# DNS name file (snippet)
myjeos IN A 192.168.1.150
racnode1 IN A 192.168.1.161
racnode1-vip IN A 192.168.1.163
racnode2 IN A 192.168.1.162
racnode2-vip IN A 192.168.1.164
rac-scan IN A 192.168.1.131
rac-scan IN A 192.168.1.132
rac-scan IN A 192.168.1.133
Or, can I just provide only the rac-scan information
# DNS name file alternate (snippet)
myjeos IN A 192.168.1.150
rac-scan IN A 192.168.1.131
rac-scan IN A 192.168.1.132
rac-scan IN A 192.168.1.133
What I am getting at is the following.
Within the install process, will racnode1, racnode1-vip, racnode2,
and racnode2-vip host names and their IP address be written
to the RAC instances /etc/hosts files? (So I should not bother
to put them in the DNS name file like '# DNS name file alternate (snippet)'?)
Or, should I put the racnode and racnode-vip host names and IP addresses
in the DNS name file like '# DNS name file (snippet)'?
The second question is the following.
Are the cluster name and the scan name allowed to be different?
Currently, I would plan them to be different,
for example, rac-cluster and rac-scan.
Or, are they required to be the same,
for example, rac-cluster and rac-cluster.
Thank you.
AIMAIM wrote:
do I need to provide the racnode1 and racnode2 information,
Or, can I just provide only the rac-scan information You need to provide all of it in DNS, because other hosts in your network will need to be able to resolve all of the normal, VIP and SCAN addresses for your RAC nodes. We write this data out to /etc/hosts just to reduce the amount of round-trip DNS requests the cluster nodes make for themselves.
Are the cluster name and the scan name allowed to be different?They can be different. -
RAC issues and mostly asked Interview questions
Hi Gurus,
It would be great if someone shares some RAC issues and mostly asked questions.
Thanks in advance.Hi,
You can also refer to the following note.
RAC and Oracle Clusterware Best Practices and Starter Kit (Platform Independent) (Doc ID 810394.1)
It has references to many useful information.
Thank you. -
How to start Oracle 10g RAC database and clusterware?
I have steps to stop the 10g RAC Database and clusterware but not sure about starting it.
I have heard executing
$crsctl stop crs --as root
on each node
will start the database,asm,nodeapps .Is that true?
or we have to do that step by step like we do in stopping the clusterware and database below
1.Stop the agent:
cd to $AGENT_HOME/corpng04.amhc.amhealthways.net/bin, then run: ./emctl stop agent
2.Stop the full database
$ oracle_home/bin/srvctl stop database -d db_name
3.Stop the ASM Instances on node1,node2
$ oracle_home/bin/srvctl stop asm -n node -- I guess you can't give multiple nodes in one command with comma,you need to give this multiple times with diff node name
4.Stop the NodeApps :vip,listener,oms and gsd
$ oracle_home/bin/srvctl stop nodeapps -n node -- I guess you can't give multiple nodes in one command with comma,you need to give this multiple times with diff node name
5.Stop the CRS cluster processes :those bloody 3 evmd,ocssd,crsd
$su - root
$CRS_home/bin/crsctl stop crsPaul R @ NL wrote:
before is shutting down crs i tend to stop the instances and services via srvctl then stop crs via crsctl
just the way i do it. not saying it's the right way but it is the one i am comfortable with.Good -) If we stop CRS, but forgot shutdown oracle instances ... we'll see shutdown abort in alert log file(that mean instances are shutdowned abort).
We should shutdown instance before stop CRS anyway. -
Compatibility with Oracle RAC 10g and E-business 11i ?
Hi netpros,
I am currently helping the sales guys with a tender. The customer requires the solution to be compatible for delivering Oracle RAC 10g and e-business 11i. we are proposing a combination of Cat6509E for the core with Infiniband server switches. Are these devices OK for use with Oracle. I have not much experience with Data centers and so any help is much appreciated.Hi,
It just so happens Cisco have a paper on exactly this subject (Oracle and e-business 11i) here:
http://www.cisco.com/application/pdf/en/us/guest/netsol/ns50/c649/ccmigration_09186a00807688ce.pdf
The design is somewhat OTT as it includes everything which Cisco thinks may be even vaguely useful, including ACE and FWSM modules. However, it's very useful as a comparison document and does include some good design tips, and the references at the end are also worth following up.
I wouldn't want to get into the whole Infiniband vs Fiber Channel argument (the doc uses MDS switches) but both work just fine in the DC environment.
HTH
Andrew.
Maybe you are looking for
-
I just bought a new iPod nano- the first I have ever owned. I installed the iTunes software and that works okay. I plugged in the nano to the computer and the computer recognizes the nano, but iTunes does not. When I open iTunes, I get the following
-
N900 October 25th update breaks music player?
I installed that latest update, Maemo 5 20.2010.36-2.002. Now the play button doesn't work on the music applet and half my songs do not appear in the music player. Does anyone else see this or know a solution? Doesn't Nokia test these updates!?
-
Oracle 9i installtion problen on RH21AS 2.4.9-e.3
Hi, I am installing Oracle 9i cluster on RH21AS kernel version 2.4.9-e.3. Using the doc "Tips to install oracle 9i on RH21AS" I have installed my OS configured it to install oracle. I could get Oracle cluster manager installed and started (using watc
-
How can I make transparent letters on a coloured background in Elements 9?
-
What is solaris 10 kernel file name ?
doing report on solaris 10 need the kernel file name