I'm a new member to this forum and I'm very new to storage too.
I have been having these errors,1370 and 1360, on SVC which is connected to DS8300 on backend; from past few months. Usually it occurs on any random mdisks and looks like a fabric problem to me.
My first query would be that what part of the whole storage is this error related to (SVC,DS8300,switch,cable,etc)?
My second query is will this error lead to a bigger problem in the future?
When contacted to IBM they said to upgrade the switches. I'm not sure how will it help but we are looking forward to do so once the client is ready.
P.S. : The SVC is on lower update because the client requires some vmware feature which is not available in furthur upgrades (We also receive the login excluded error due to some linux kernel issue). Upgrade is impossible without that feature.
We have two brocade 5300 Switches at 6.2.0e connected to the SVC as well as IBM DS8300.PFA
Pinned topic SCSI ERP occurred on SVC (18.104.22.168)
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2013-01-15T10:41:09Z at 2013-01-15T10:41:09Z by seb_
SystemAdmin 110000D4XK4779 Posts
Re: SCSI ERP occurred on SVC (22.214.171.124)2013-01-15T06:50:22ZThis is the accepted answer. This is the accepted answer.Hi Kamal,
this event means that SVC is receiving SCSI check condition from the device.
It's not possible to say more using only event code 1370. For detailed investigation error data should be decoded.
Actually this event indicates the problem which is most likely outside the SVC itself. It can be a cable, backend storage, or even code compatibility problem.
So the best way here is to stick to IBM's recommendations regarding switches, as IBM service can decode event data and they know what's going on.
seb_ 060000QVK225 Posts
Re: SCSI ERP occurred on SVC (126.96.36.199)2013-01-15T10:41:09ZThis is the accepted answer. This is the accepted answer.As you have 1360s, too, the 1370s are most probably "follow-on-symptoms" of problems in the fabric. As Sergey Kubin explains they could contain sense data from check condition,, meaning the backend storage as a target wanted to make the SVC as an initiator aware of a special situation. The 1360s are most probably indications of a physical problem or a bottleneck in your case. I cannot track it down from the picture, but the errors themselves should contain more info. The 1370s could then mean that there were PORs, for example because there were LUN resets or target resets subsequently.
Here are some hints how to read 1370 (SCSI ERP occured): https://www.ibm.com/developerworks/mydeveloperworks/blogs/sanblog/entry/don_t_shoot_the_messenger_the_error_13706