FICON would help with the data transfer, but PAV is what would help with the
multiple simultaneous reads/writes.

Since you are ESCON, you are not PAV capable, so the next best solution
would be to spread the data across as many volumes as possible.  If you have
a control unit capable of creating many 'small' 3390 images rather than
fewer 'big' images, then
odds are your access profile will be improved at the volume level by using
many small volumes.  'Small' and 'big' depend on the application, but for
example, you would generally be better off from the DASD performance
perspective to spread data over 108 3390-1 volumes than to put it on 12
3390-9 volumes.

FYI, IBM has a document addressing Linux and PAV.

http://oss.software.ibm.com/linux390/docu/lx24pav00.pdf


Scott Ledbetter
StorageTek



-----Original Message-----
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Marcy
Cortes
Sent: Thursday, July 15, 2004 10:17 AM
To: [EMAIL PROTECTED]
Subject: Re: Performance of large file systems


You might want to see if you can get FICON instead of ESCON.


Marcy Cortes

This message may contain confidential and/or privileged information.  If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose, or take any action based on this message or
any information herein.  If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message. Thank
you for your cooperation."

-----Original Message-----
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Thomas
Denier
Sent: Thursday, July 15, 2004 07:21
To: [EMAIL PROTECTED]
Subject: [LINUX-390] Performance of large file systems

We are considerng a file organization with rather unusual characteristics,
and are wondering if we are likely to run into performance pathologies.

The proposed file organization would be used to provide a TSM server with a
family of sequential storage pools sharing a single device class with
DEVTYPE=FILE.

We would use LVM to assemble a large number of ESCON 3390 volumes into a
single file system with a size in the hundreds of gigabytes. This would be
an ext3 file system with metadata journalling.

All files in the file system would reside in a single directory.

The directory would start out empty each night, and end up with a few
hundred files. Some of the files would be as small as tens of megabytes,
some would be two gigabytes, and some would have various sizes in between.

There would sometimes be as many as seventy files open for writing at the
same time. All writes would be appends. There would sometimes be as many as
four files open for reading while the write activity was going on. No
individual file would be open for reading and writing at the same time.

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions, send email
to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions, send email
to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to