Unsolved
This post is more than 5 years old
3 Posts
0
3154
storapid takes wrong disk - not the gatekeeper
My storapid process on AIX takes the wrong hdiskpower for his communications.
For example:
# lspv | grep hdiskpower
hdiskpower0 00c8e5aa40c15065 dhb
hdiskpower1 00c8e59aafa5cd2e dat0vg active
hdiskpower2 00c8e5aaaf9dc184 applvg
hdiskpower3 none None
hdiskpower4 none None
hdiskpower5 none None
hdiskpower6 none None
# symgate list
( ... )
Device Name Directors Device
--------------------------- ------------- -------------------------------------
Cap
Physical Sym SA DA :IT Config Attribute Sts (MB)
--------------------------- ------------- -------------------------------------
/dev/rhdiskpower3 0068 10A:1 16B:CE 2-Way Mir N/Asst'd GK RW 3
/dev/rhdiskpower4 0069 10A:1 01C:CE 2-Way Mir N/Asst'd GK RW 3
/dev/rhdiskpower5 006A 10A:1 01D:DE 2-Way Mir N/Asst'd GK RW 3
/dev/rhdiskpower6 006B 10A:1 16A:CF 2-Way Mir N/Asst'd GK RW 3
lsof show that hdiskpower0 and hdiskpower3 are in use by the storapid process.
Because of that I cannot use hdiskpower0 for my AIX purposes.
Versions:
AIX :5300-11-01-0944
powerpath: 5.3.1.0
Solutions enabler:7.1.2.0
Someone an idea ?
Nollaig1
1 Rookie
1 Rookie
•
137 Posts
0
January 13th, 2011 03:00
Hello Rob,
Looking through the knowledgebase, article emc255976 is very informative in regards gatekeepers in general, and article emc186179 verifies that gatekeepers are single pathed but what you have listed below indicates that anyway.
Can you tell me is this a new install? Has this host previously used these gatekeepers?
Have you checked that the GK are accessible from the host?
Anything in the errprt?
And would you have any output from the symapi log file (or all daemon log files) as well please?
Have you tried stopping and starting the daemon?
Sorry for the list of questions.
Thank you Rob,
Nollaig
1RobR
3 Posts
0
January 14th, 2011 06:00
Hello Nollaig,
Thanks for your review.
Yesterday I got the answer from EMC support ( I had to make a problem report because of time issues )
EMC gave me the solution: here it is:
The disk hdiskpower0 comes from another symmetrix( let's call it the remote symmextrix ).
It is used as a heartbeat disk between this node and his 'brother'-node.
hdiskpower1 and hdiskpower2 are from the local EMC symmetrix.
So, there is also sym-communication needed between this node and the remote EMC symmetrix.
But there are not any Gatekeeper devices from the remote symmetrix and Sol.enabler takes hdiskpower0 for his communication.
So we added a few gatekeepers from the remote symmetrix and the problem is solved !
I now have 2 GK devices from the remote symmetrix and 4 from the local symmetrix.
EMC advice is to have 6 GK devices.
In my situation then I will have a total of 12 GK devices.
Unfortunately this gives a big output of lspv , but that is only the output, nothing harming.
Hope you learn from my text and thanks again !
Rob
SKT2
1.3K Posts
0
January 29th, 2011 18:00
That is interesting, never heard of a remote GK usage. Did EMC mention about any primus being updated/existed with this information?
1RobR
3 Posts
0
January 29th, 2011 19:00
Ik ben niet aanwezig vanaf 12-11-2010 en kom terug op 01-01-2016.
Geachte, deze mailbox is niet meer in gebruik. Uw mail zal dan ook niet
gelezen en beantwoord worden. U kunt uw mail doorsturen naar
1Rob_Roos@nl.ibm.com