Packet errors on all isci interfaces after upgrading to 6.0.1
After upgrading to 6.0.1 firmware on one of my member arrays, the array is constantly generating packet errors and an amber warning status on the interfaces in the group manager interface. I have a 3 member group with two running 5.2.2 and one running 6.0.1. Before I upgrade the other two, I would like to know if anyone else has seen this behavior. The switches I am using are 2 Dell 5548 stacked, and I have never seen this issue before on any other version of firmware.
I am seeing the same issue when I upgrade a PS4100X from 5.2.5 to 6.0.1. Restarted multiple times to toggle active controller, shutdown and powered off multiple times. Even reset unit and reconfigured. iSCSI interfaces still have packet errors, at least 20/minute. Unit was working perfectly beforehand.
Upgraded PS4000X from 5.2.5 to 6.0.1, no issues so far, 4 days running.
DCB was enabled. I disabled it and saved. Still no difference. Lots of errors. Restarted, still errors.
The 4100X as well as the 4000X are connected to two PowerConnect 6224's connected to each other with stacking cables. Again, the 4000X has no errors, the 4100X has thousands of errors in under an hour.
I had a support case open with Dell regarding this. All of my members (3 in total and a mix of ps6000s and ps4100) were experiencing the packet error behavior after upgrading to 6.0.1. They had me disable STP on the switch ports the arrays were plugging into directly and that seems to have fixed the issue.
Thanks for the replies. Apparently over the course of a year, the SAN cables had been plugged into ports that didn't have STP disabled. Once I moved the cables back to the original ports, everything was clean. I verified by then plugging into different ports, and had the problem again. I then disabled STP on those ports, and everything was clean again.
So in my case, STP seems to have been the issue. Thanks again for your replies.
I would have opened up a ticket with Dell support, but I was already dealing with IBM support on an unrelated issue. This wasn't critical yet, and would have been escalated to support once the IBM issue was fixed.
We have same issue, on 7.0.9, STP on as best practices, so... is this going to be fixed or should we disable STP? See no mention in post hotfixes, support sent their canned upload logs email, etc... Not the best customer service. We also see packet errors on management port. The LAN switch also has STP on. Don't want to take step backwards. Switches do not show errors, only within the GPADMIN portal.
Sorry, actually meant spanning tree portfast. We do use Dell switches. So we need to remove spanning-tree portfast for all connections to/from the SAN to iSCSI switch? including the management port to LAN switch?
laserdude69
3 Posts
984
0
Posted October 5th, 2012 18:00
I am seeing the same issue when I upgrade a PS4100X from 5.2.5 to 6.0.1. Restarted multiple times to toggle active controller, shutdown and powered off multiple times. Even reset unit and reconfigured. iSCSI interfaces still have packet errors, at least 20/minute. Unit was working perfectly beforehand.
Upgraded PS4000X from 5.2.5 to 6.0.1, no issues so far, 4 days running.