Your test / complaint I believe is about the start up time for an z/OS sshd
session.

You may wish to enable:

LOGLEVEL DEBUG3

tracing for SSHD in sshd_config (to syslogd) .
This would be helpful to see where all of the time is going.

A couple of wild guesses:

- Are you starting SSHD with INETD?   It is much quicker to start its own
SSHD proc

- You don't have ICSF configured to generate random numbers.
  See "Using hardware support to generate random numbers" in the IBM Ported
Tools UG.
  NOTE:  with ICSF HCR77A0, you don't need a crypto-card for ICSF
/dev/random support.

  Anyone who is still using the crappy ssh-rand-helper instead of ICSF
/dev/random should look into HCR77A0, which is MUCH better.
  ssh-rand-helper is terrible - slow and expensive and not as secure.



Kirk Wolf
Dovetailed Technologies
http://dovetail.com


On Mon, Jul 8, 2013 at 9:16 AM, David Crayford <[email protected]> wrote:

> IIRC, IBM have a new release of SSH that has CPACF acceleration.
>
>
> On 8/07/2013 9:57 PM, Paul Gilmartin wrote:
>
>> Comparing performance of ssh to MVS and Solaris severs, respectively:
>>
>>  From Solaris to MVS:
>>
>> 133$ time ssh user@MVS         date
>> Mon Jul  8 07:43:06 MDT 2013
>>
>> real    0m15.10s
>> user    0m0.07s
>> sys     0m0.01s
>>
>>  From Solaris to another Solaris:
>>
>> 134$ time ssh user@solaris     date
>> Monday, July  8, 2013 07:43:57 AM MDT
>>
>> real    0m0.61s
>> user    0m0.15s
>> sys     0m0.01s
>>
>> The MVS performance is awful (in the synchronic sense).
>> Is there any way to tell where the overhead lies, or
>> even whether ICSF is being used rather than ssh_rand_helper?
>>
>> How does this compare with other users' experience?
>>
>> (Once an interactive connection is established, response
>> is quite good.)
>>
>> -- gil
>>
>> ------------------------------**------------------------------**
>> ----------
>> For IBM-MAIN subscribe / signoff / archive access instructions,
>> send email to [email protected] with the message: INFO IBM-MAIN
>>
>
> ------------------------------**------------------------------**----------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN
>

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to