Re: iscsi problem for two nics

2009-09-10 Thread Yao Wei
In short: Case A Initiator: IBM x3550 (two NICs), /dev/sda (local disk), 100-120 MB/s (hdparm -t /dev/sda). Target:Dell (only one NIC), /dev/sda (local disk), dev/sda (IET disk),70-90 MB/s(hdparm -t

8 monkeys

2009-09-10 Thread Nikhil Talpallikar
(This is reportedly based on an actual experiment conducted in the U.K.) * * *Put eight monkeys in a room. In the middle of the room is a ladder, leading to a bunch of bananas hanging from a hook on the ceiling.* * * * * *Each time a monkey tries to climb the ladder, all the monkeys are sprayed

Re: iscsi problem for two nics

2009-09-10 Thread Paul
has anybody tested this with 10Gb cards? what is the performance gains? I have tested it lightly and found that there are some gains but without a 10Gb switch in the mix testing is very difficult. Paul Yao Wei wrote: In short:

Question about iSCSI Enterprise Target

2009-09-10 Thread Michael Schwartzkopff
Hi, perhaps this is the wrong mailing list, but anyway, perhaps anybody here can help me: iSCSI Enterprise Target is the projekt for the Linux implementation of a iSCSI target. My question is wheather the software can be used to bind the same target to two different initiators? Thanks for

Re: Question about iSCSI Enterprise Target

2009-09-10 Thread Pasi Kärkkäinen
On Tue, Sep 08, 2009 at 08:14:38PM +0200, Michael Schwartzkopff wrote: Hi, perhaps this is the wrong mailing list, but anyway, perhaps anybody here can help me: iSCSI Enterprise Target is the projekt for the Linux implementation of a iSCSI target. My question is wheather the

Re: [PATCH 1/4] libiscsi: Check TMF state before sending PDU

2009-09-10 Thread Mike Christie
On 08/13/2009 02:58 AM, Hannes Reinecke wrote: Before we're trying to send a PDU we have to check whether a TMF is active. If so and if the PDU will be affected by the TMF we should allow only Data-out PDUs to be sent, or, if fast_abort is set, no PDUs at all. Hey, I updated this patch with

monitoring iscsi connections

2009-09-10 Thread solidguy
Is there a good way to monitor the iscsi sessions for errors? I need to report it via SNMP (and perhaps kill the VM running over the connection). Right now the only way I see is to monitor syslog -- but the syslog doesn't always give me which session (or disk) is in trouble. Thanks -- solidguy

Re: monitoring iscsi connections

2009-09-10 Thread Mike Christie
On 09/10/2009 12:53 PM, solidguy wrote: Is there a good way to monitor the iscsi sessions for errors? I need to report it via SNMP (and perhaps kill the VM running over the connection). Right now the only way I see is to monitor syslog -- but the syslog doesn't always give me which session

[PATCH] add target reset support

2009-09-10 Thread Mike Christie
This patch adds target reset support. If your driver was already setting the sht-eh_target_reset_handler callout to iscsi_eh_target_reset you do not have to make any code changes. The old iscsi_eh_target_reset function was just dropping the session. Now it will try a warm target reset. If that

Re: monitoring iscsi connections

2009-09-10 Thread Chiradeep Vittal
Thanks. I'll take a look at the netlink interface. Not using multipath for now, but will do so later. For basic monitoring of storage network problems, here's what I am thinking: 1. If there is a network failure, eventually cat /sys/block/disk/ device/state should show offline ? 2. How long will

Apologies!!!

2009-09-10 Thread Nikhil Talpallikar
Hi, Very sorry for the previous mail which was actually intended for a few people but somehow, by mistake got to all of the contacts and other mailing lists I had ever sent a mail to. I will take care that such a human error won't repeat again. Apologies again!! Thanks and Regards, Nikhil