Question : Problem: WTF's up with my iSCSI network config ???

Experts:

I just purchased an EMC AX4-5i dual-SP SAN appliance; two racks, one for SAS drives and the other with SATA drives. I'm just setting up the appliance and I'm stuck, hoping you all can help me figure something out.

If you look at the attached file you'll notice my vanilla setup: 1 server with 3 NICs connected to a pair of GigE switches configured in a meshed network connecting a pair of SP units, each with two iSCSI ports of their own.

The problem I'm having is that on the server i can only ping one of two switches and only two of four iSCSI ports

C:\Program Files\Support Tools>ipconfig /all

Windows IP Configuration

   Host Name . . . . . . . . . . . . : galapagos
   Primary Dns Suffix  . . . . . . . : xxx.local
   Node Type . . . . . . . . . . . . : Unknown
   IP Routing Enabled. . . . . . . . : Yes
   WINS Proxy Enabled. . . . . . . . : Yes
   DNS Suffix Search List. . . . . . : xxx.local


Ethernet adapter 192.168.253.98:

   Connection-specific DNS Suffix  . : xxx.local
   Description . . . . . . . . . . . : Intel(R) PRO/1000 MT Server Adapter
   Physical Address. . . . . . . . . : 00-04-23-AB-6A-0B
   DHCP Enabled. . . . . . . . . . . : No
   IP Address. . . . . . . . . . . . : 192.168.253.98
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Default Gateway . . . . . . . . . :

Ethernet adapter 192.168.253.99:

   Connection-specific DNS Suffix  . :
   Description . . . . . . . . . . . : Intel(R) PRO/1000 MT Server Adapter #2
   Physical Address. . . . . . . . . : 00-04-23-AB-6A-0C
   DHCP Enabled. . . . . . . . . . . : No
   IP Address. . . . . . . . . . . . : 192.168.253.99
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Default Gateway . . . . . . . . . :

Ethernet adapter 192.168.10.25:

   Connection-specific DNS Suffix  . :
   Description . . . . . . . . . . . : HP NC3163 Fast Ethernet NIC
   Physical Address. . . . . . . . . : 00-50-8B-EB-15-1C
   DHCP Enabled. . . . . . . . . . . : No
   IP Address. . . . . . . . . . . . : 192.168.10.25
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Default Gateway . . . . . . . . . : 192.168.10.1
   DNS Servers . . . . . . . . . . . : 192.168.10.13
                                       192.168.10.25
   Primary WINS Server . . . . . . . : 192.168.10.13
   Secondary WINS Server . . . . . . : 192.168.10.25

C:\Program Files\Support Tools>ping 192.168.253.199

Pinging 192.168.253.199 with 32 bytes of data:

Reply from 192.168.253.199: bytes=32 time=3ms TTL=64
Reply from 192.168.253.199: bytes=32 time=2ms TTL=64
Reply from 192.168.253.199: bytes=32 time=1ms TTL=64
Reply from 192.168.253.199: bytes=32 time=2ms TTL=64

Ping statistics for 192.168.253.199:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 1ms, Maximum = 3ms, Average = 2ms

C:\Program Files\Support Tools>ping 192.168.253.198

Pinging 192.168.253.198 with 32 bytes of data:

Request timed out.
Request timed out.
Request timed out.
Request timed out.

Ping statistics for 192.168.253.198:
    Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

C:\Program Files\Support Tools>ping 192.168.253.200

Pinging 192.168.253.200 with 32 bytes of data:

Reply from 192.168.253.200: bytes=32 time<1ms TTL=64
Reply from 192.168.253.200: bytes=32 time<1ms TTL=64
Reply from 192.168.253.200: bytes=32 time<1ms TTL=64
Reply from 192.168.253.200: bytes=32 time<1ms TTL=64

Ping statistics for 192.168.253.200:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 0ms, Maximum = 0ms, Average = 0ms

C:\Program Files\Support Tools>ping 192.168.253.201

Pinging 192.168.253.201 with 32 bytes of data:

Request timed out.
Request timed out.
Request timed out.
Request timed out.

Ping statistics for 192.168.253.201:
    Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

C:\Program Files\Support Tools>ping 192.168.253.202

Pinging 192.168.253.202 with 32 bytes of data:

Reply from 192.168.253.202: bytes=32 time=1ms TTL=64
Reply from 192.168.253.202: bytes=32 time<1ms TTL=64
Reply from 192.168.253.202: bytes=32 time<1ms TTL=64
Reply from 192.168.253.202: bytes=32 time<1ms TTL=64

Ping statistics for 192.168.253.202:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 0ms, Maximum = 1ms, Average = 0ms

C:\Program Files\Support Tools>ping 192.168.253.203

Pinging 192.168.253.203 with 32 bytes of data:

Request timed out.
Request timed out.
Request timed out.
Request timed out.

Ping statistics for 192.168.253.203:
    Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

C:\Program Files\Support Tools>

So before I go any further and start configuring iSCSI initiators or LUNs, etc...I wanted to clear up this networking mystery


Thanks,
juckyt

Answer : Problem: WTF's up with my iSCSI network config ???

So you are using DTS and not a third party ETL tool?  My first suggestion, with load times getting that high, is start the path to an ETL tool (informatica, ascential, etc).

Short of that, we need to know more about what the steps in your jobs are doing--updates will always be a killer (especially without indexes--if you dropped them all).  Inserts aren't so bad--especially if you tell DTS to lock the table.  Also, clustered indexes--we learned that on large data stores--you never drop the clustered indexes--the time to rebuild it at the end offsets the benefits.

Use ActiveX scripts as little as possible--the overhead is substantial.  Use data driven query tasks and lookups as little as possible--also lots of overhead.  The more "manipulation" you can perform with SQL usually the better--let the database engine do a lot of the work and not DTSRUN.

Of course, if you're running RAID (I would assume you are) and you can afford it--use RAID10--the write performance on a data load is AWESOME.  Also, make sure you have the database log on different disks/RAID groups.  If you can get tempdb on a different drive/raid group even better......

Splitting across multiple files (filegroups) will buy you more sql worker threads too.  You figure you get at least 1 thread per file (minimum).....

The above are just a few things.  If you describe more about steps and what they do, we might be able to offer more assistance.

What store does the data come from?  Have you isolated your longest running steps?

brett