Re: ISCSI EMC OEL performance issue
- From: Gaja Krishna Vaidyanatha <gajav@xxxxxxxxx>
- To: "oracle-l@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
- Date: Wed, 7 Mar 2012 10:17:05 -0800 (PST)
Hi Vamshi,
Hope you are doing well. Storage Performance Troubleshooting in a SAN is like
unraveling a 7-layer burrito....:) There are many layers of abstraction in a
SAN and the concept of "throughput measurement" (in your case data transfer
rates) has to be done keeping the "lowest common denominator". What I mean here
is this - Even if your storage volume can push 1024MB/sec in data transfers,
your net throughput can be limited by the other parts in your SAN (i.e.
switches/HBAs/bus etc). If your switch/HBA supports 6Gb/sec, the maximum
theoretical throughput you can achieve is 768 MB/sec. This is because the "b"
in "6Gb" is a small b, standing for Gigabit, so one has to do the relevant
arithmetic (divide by 8) for this. I pulled up the specs for your storage array
- http://www.emc.com/collateral/software/specification-sheet/h8514-vnx-series-ss.pdf,
but it does not quite tell me what this beast can actually achieve from a data
transfer standpoint. May be I have
missed something in the spec. It may not be a bad idea to sit down with your
EMC storage folks and the internal storage team, layout the physical
architecture and tag some numbers on the data transfer rates of each component.
This will then give you the true theoretical limit for data transfer. And that
could be a better starting point to set expectations on what your SAN can
deliver. Let me know if you have any further questions.
Cheers,
Gaja
Gaja Krishna Vaidyanatha,
CEO & Founder, DBPerfMan LLC
http://www.dbperfman.com
http://www.dbcloudman.com
Phone - +1-650-743-6060
http://www.linkedin.com/in/gajakrishnavaidyanathaCo-author:Oracle
Insights:Tales of the Oak Table
- http://www.apress.com/book/bookDisplay.html?bID14
Co-author:Oracle Performance Tuning 101
- http://www.amazon.com/gp/reader/0072131454/ref=sib_dp_pt/102-6130796-4625766
Enabling Cloud Deployment & Management for Oracle Databases
________________________________
From: Vamshi Damidi <dbaprimatics@xxxxxxxxx>
To: oracle-l@xxxxxxxxxxxxx
Sent: Tuesday, March 6, 2012 8:19 AM
Subject: ISCSI EMC OEL performance issue
Hi All,
Hope every one is doing great, busy with research.
Any suggestions to this post is appreciated.
DESC: We have bought new SAN EMC VNX 5300 / OEL / Broadcom CNA 10G / ISCI
Lun Size 5**G
Disk : 15k RPM * 13 disks
Current IO calibaration results :
START_TIME END_TIME MAX_IOPS
MAX_MBPS MAX_PMBPS LATENCY NUM_PHYSICAL_DISKS
------------------------------ ------------------------------ ----------
---------- ---------- ---------- ------------------
28-FEB-12 03.26.15.587453 PM 28-FEB-12 03.33.53.600931 PM
4148 433 295 11 24
i observe that actual is not match or near to calculated.
Can you please suggest some thoughts for imporvment.
[root@srvpdb01 ~]# cat /etc/iscsi/iscsid.conf
node.startup = automatic
node.session.timeo.replacement_timeout = 120
node.conn[0].timeo.login_timeout = 15
node.conn[0].timeo.logout_timeout = 15
node.conn[0].timeo.noop_out_interval = 10
node.conn[0].timeo.noop_out_timeout = 15
node.session.initial_login_retry_max = 4
node.session.cmds_max = 128
node.session.queue_depth = 128
node.session.iscsi.InitialR2T = No
node.session.iscsi.ImmediateData = Yes
node.session.iscsi.FirstBurstLength = 16776192
node.session.iscsi.MaxBurstLength = 16776192
# the default is 131072
node.conn[0].iscsi.MaxRecvDataSegmentLength = 16776192
# the default is 32768
discovery.sendtargets.iscsi.MaxRecvDataSegmentLength = 16776192
node.conn[0].iscsi.HeaderDigest = None
node.session.iscsi.FastAbort = No
net.core.rmem_max = 1073741824
net.core.wmem_max = 1073741824
net.ipv4.tcp_rmem = 1048576 16777216 1073741824
net.ipv4.tcp_wmem = 1048576 16770216 1073741824
net.ipv4.tcp_mem = 1048576 16770216 1073741824
Above are my current operational parameters.
Thanks,
Vamshi .D
--
//www.freelists.org/webpage/oracle-l
--
//www.freelists.org/webpage/oracle-l
Other related posts:
