How to find the no. of cores in a Solaris box

I have finally found how to find the no. of cores in a Solaris Box on newer T class servers. The clue was given on Steve Sistare's blog post CPU to Core Mapping. He mentioned that:

"psrinfo -pv" only prints the core information on systems running OpenSolaris or Solaris Express, because psrinfo was enhanced by this CR:

6316187 Need interface to determine core sharing by CPUs

which was never backported to a Solaris 10 update.

So I googled for psrinfo opensolaris and found this perl script.

Here's the ouptut of running this perl script on a T5220 server:

$ perl psrinfo.pl -pv

The physical processor has 4 cores and 13 virtual processors (-50)

Another command to get the number of cores is:

kstat cpu_infogrep core_idsort -uwc -l

The core has 4 virtual processors (0-3)

The core has 3 virtual processors (16-18)

The core has 3 virtual processors (32-34)

The core has 3 virtual processors (48-50)

UltraSPARC-T2 (clock 1415 MHz)

Warning: log write time

After upgrading to 10.2.0.4, we started seeing messages in alert log

*** 2009-10-27 11:12:26.575
Warning: log write time 590ms, size 22KB

Metalink Note 601316.1 describes this issue. LGWR starts producing trace files with this kind of information:

Elapsed times include waiting on following events:

Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
SQL*Net message to client 3 0.00 0.00
SQL*Net message from client 3 0.00 0.00
db file sequential read 7080 3.83 196.43
latch: session allocation 1 0.00 0.00
read by other session 2 0.53 0.59
latch free 1 0.00 0.00

*** 2009-10-27 11:12:26.575
Warning: log write time 590ms, size 22KB
*** 2009-10-27 11:12:28.294
Warning: log write time 610ms, size 43KB
*** 2009-10-27 11:12:30.257
Warning: log write time 570ms, size 69KB
*** 2009-10-27 11:12:47.335
Warning: log write time 700ms, size 2213KB
*** 2009-10-27 11:14:02.255
Warning: log write time 560ms, size 1372KB
*** 2009-10-27 11:14:03.104
Warning: log write time 850ms, size 2823KB
*** 2009-10-27 11:16:10.756
Warning: log write time 640ms, size 261KB
*** 2009-10-27 11:16:51.149
Warning: log write time 500ms, size 1703KB
*** 2009-10-27 11:16:51.800
Warning: log write time 650ms, size 3317KB

As per the note

These messages are very much expected in 10.2.0.4 database in case the log write is more than 500 ms.

This is a warning which means that the write process is not as fast as it intented to be. So probably you need to check if the disk is slow or not or for any potential OS causes. If everything looks fine at the hardware level or OS level then you can safely ignore these messages.

The trace file can easily be deleted or truncated.

How to download password protected patch from updates.oracle.com

Akhilesh asked me this question today. A metalink search revealed note 731257.1 which has the details. In short before issuing the get command you are required to issue the command site protected .

In fact if you directly issue the get command you get the error message which says the same:

ftp> bin
g200 TYPE changed to I.
ftp> et p8995444_11i_d.zip
200 PORT command OK.
550-File or directory not found.
550-This is a password protected patch.
550-Enter the password with site command:
550-quote site protected
550 Then get the patch again.

After issuing the site protected command it worked

ftp> site protected V5pIHQqk
200 The password was received successfully.
ftp> get p8995444_11i_d.zip
200 PORT command OK.
150 Opening BINARY mode data connection for file p8995444_11i_d.zip.

It worked with quote site protected command also:

ftp> quote site protected V5pIHQqk
200 The password was received successfully.
ftp> get p8995444_11i_d.zip
200 PORT command OK.
150 Opening BINARY mode data connection for file p8995444_11i_d.zip.

This is of limited use now as Oracle is decomissioning their ftp service and you'll need to use wget and URL from the patch download button in new metalink.