Solved!
Go to Solution"Unit Shutdown for Trespass"
we've seen a lot of "Unit Shutdown for Trespass" in the SP event log, it logs a entry like in about every 5min, see attached screen print, the luns are part of RLP luns, is this a normal activity? I can understand sometimes a lun is trespassed… but every 5 minutes…? that doesn’t sound right. this is on a CX3 with MV/A. thanks
2238
0
Posted August 16th, 2010 11:00
thank you both for the info!
2238
1
Posted August 16th, 2010 11:00
RLP LUNs when not assigned to a session are not "owned" by either SP. When a snap session or M/A session first starts it is assigned an RLP and there is an initial trespass to get the LUN to the correct SP - this is normal. You may have processes running that are grabbing a RLP and then releasing it. This could account for the messages.
This is one of the reasons the are some restrictions on where you create the RLPs - you should have a dedicated Raid Group (s) just for RLP and you should not have any normal LUNs in these raid groups and you should not use the first five disks in the array (Vault Drives).
If you're seeing a performance issue, you should probably open a case wtih EMC. but if you have the RLP configured as above, you should not see issues unless the RLP are using SATA (not recommended) or the source LUNs are using faster disks or more disks than the RLP raid groups.
glen
2238
0
Posted August 16th, 2010 15:00
Lennie, are you sure these are RLP luns that are trespassing....
There is a BV running on this lun that is getting trespassed. I am not sure that BV runs on RLP luns.
Do you have any user lun in the raid group, that disk being called out is in?
disk is in 5_0 (but cannot see what the actual disk is).
2238
0
Posted August 16th, 2010 15:00
Forgot to add that the RLP luns will trespass also if the lun it is attached to (source lun) also trespasses.
2238
1
Posted August 18th, 2010 07:00
Hi,
Also you can refer primus article emc88102 .
2238
0
Posted August 18th, 2010 07:00
thanks for looking into this for me...
what i did was coverting the hex code of Ext Code2 to a lun number 170 (hope i did this right) which is a RLP LUN
Event Code:0x60a
Description:Internal information only. A logical unit has been enabled
Subsystem:APM00064203685
Device:Enclosure 5 Disk 0
SP:SPA
Host:spa-3685
Source:N/A
Category:N/A
Log:Storage Array
Sense Key:0x0
Ext Code1:0xc0030
Ext Code2:0xaa004b
Type:Information2238
1
Posted August 18th, 2010 18:00
Lennie, a quick check is to see if the RLP lun you think it is , has disk 0-5-0 as one of its disks....
but your conversion looks correct...
RG 12 (0x000c)
lun 170 (0x00aa)
Jim
RyanP2
261 Posts
2238
0
Posted August 16th, 2010 11:00
I have seen this before, and yes it is normal from what I have heard. The RLPs are part of a global pool on the array which means that any RLP lun can be used for any MV/A session. When the session needs to grab a lun for the snapshot it takes, the RLP lun can be any lun, which means it could be on either SP. I don't think the system has an alogithm that says to choose a lun on that SP before pulling one from the other. I also don't think there is any way to avoid this trespassing.
-Ryan