First, if anyone is running the latest NexentaStor (Community or Enterprise Editions) in production, I'd love to hear from them. It seems like a fairly robust appliance OS that provides a lot of bang for the buck. However, I am seeing tons of "overflow" warrnings in my Vmware logs. Wondering if anyone else is seeing this?
I am doing some testing of the NexentaStor OpenSolaris Storage Appliance. I have installed it on some physical hardware in my lab and am specifically testing the iSCSI Target functionality. I am receiving the following errors in my Vmware Vmkernel log:
*Jul 9 14:30:21 esx-2 vmkernel: 0:18:52:46.487 cpu6:4256)WARNING:
iscsi_vmk: iscsivmk_ConnCommandResponse: SCSI response overflow:
residual 512*
*Jul 9 14:30:21 esx-2 vmkernel: 0:18:52:46.487 cpu6:4256)WARNING:
iscsi_vmk: iscsivmk_ConnCommandResponse: Sess [ISID: 00023d000001
TARGET: iqn.1986-03.com.sun:02:164fd6ca-c6d3-e437-aafd-d51ee7beeaef
TPGT: 2 TSIH: 0]*
*Jul 9 14:30:21 esx-2 vmkernel: 0:18:52:46.487 cpu6:4256)WARNING:
iscsi_vmk: iscsivmk_ConnCommandResponse: Conn [CID: 0 L:
10.0.0.101:54764 R: 10.0.0.3:3260]*
*Jul 9 14:30:21 esx-2 vmkernel: 0:18:52:46.488 cpu6:4256)WARNING:
iscsi_vmk: iscsivmk_ConnCommandResponse: SCSI response overflow:
residual 512*
Filed an SR w/ Vmware support and they responded:
+"My name is Siddharth and I am assuming ownership of your Support
Request, SR# 1544428471, and will work with you to resolve this issue.+
As I understand from the Support Request notes, the issue you are experiencing is "warning in vmkernel logs".
As you mentioned you are looking information regarding "SCSI response overflow: residual 1024"
There are 2 parts of this Scsi message
+1) Response: Response code contains a code that indicates the presence
or absence a iScsi protocol errors, The scsi response code is directed
to Scsi delivery subsystem what the Scsi status cose is to the SAL.+
+2) Overflow: When a bidirectional/Unidirectional command generates more
data than it is expected by the initiator, this bit is set to 1 and is
reflected as Bi/Unidirectional Read residual overflow.+
Summary:
+The overflow/underflow is determined by the SCSI layer. This is because
the "expected data transfer length" gives the size of the initiators
buffer and the CDB gives the size of the "attempted transfer". If the
target is attempting to transfer too much, that would be an overflow.
In some transports (like parallel SCSI) this is not determined by the
target but is determined by the initiator driver or host adapter. In
this case the target will try to stay in data-in phase to transfer all
of the data. The initiator sees this as an overflow.+
+As iScsi layer on the host is responding to the data received from the
target with more than what is requested is the result of this warning
message This is just a warning message and can be safely ignored as
normal operation will proceed and excess data will be truncated"+
I have taken the response that Vmware Support provided back to NexentaStor support, who finally after 3 weeks took me up on the offer of a packet trace, but they have not responded w/ anything useful.
Anyone else seeing this w/ NexentaStor?
No, but I am seeing the same thing with OpenSolaris snv_134. I have not determined yet whether it's really a problem or not.
I would imagine asking on one of the OpenSolaris mailing lists would be a better avenue for support than asking Nexenta.