Unsolved
This post is more than 5 years old
1 Message
0
7994
iscsi dropping when trying to move anything greater than 20gb of data
Hi Guys,
i been having some iscsi dropping when trying to move significant amount of data into the storage. Windows just dropped the drive and nothing can be done until the system is rebooted.
i got an equallogic ps6010 and R710 running windows r2 enterprise 64bits.
please assist me in finding a solution to this problem.
thanks,
bc_nguyen99
mfkelly_tx
31 Posts
0
July 12th, 2011 05:00
Do you mean moving it from one storage pool to another? 20GB is not very much, I move 2TB volumes to different storage pools all the time with no problem. Do you have the latest HIT kit installed?
DELL-Joe S
7 Technologist
7 Technologist
•
729 Posts
0
July 12th, 2011 05:00
bc_nguyen99,
Ensure your servers are fully patched, and you are running the latest versions of the iSCSI initiator. If using a HBA ensure the drivers and boot code are the latest versions.
Review the host server(s) event logs to see what events took place at the time of the iSCSI disconnect.
Look at the Array event log, to see if you see anything unusual there.
You also need to ensure that your SAN network is setup properly for high performance for iSCSI traffic (flow control enabled end to end). Also ensure that your iSCSI connections are not on the default VLAN, there should be a separate vlan for iSCSI even if the switch is dedicated to iSCSI.
Dell EqualLogic Network Performance Guidelines v4.0: attachments.wetpaintserv.us/CjSA1VnfNTHzy7O03UbN0g485148
Also ensure that the iSCSI disk timeouts are setup properly on you host servers. On the firmware download page, you need to setup per the “iSCSI Initiator and Operating System Considerations” link
Lastly, ensure your EqualLogic arrays are on the latest version of the firmware.
Regards,
Joe
evilensky
74 Posts
0
July 12th, 2011 13:00
We've had no end of trouble with Broadcom chipsets on 2008 R2 and our OS build; we've made it a practice to stabilize systems by turning off: Receiver Side Scaling, TCP Chimney Offload, and NetDMA. You're mileage may vary, but I do recommend try disabling these settings to see if it makes a difference. For us it became the difference between a perfectly stable system (but potentially not 100% optimized) and one that rebooted whenever heavy SQL IO appeared.