/dev/async (387 Views)
Reply
Occasional Advisor
Deepanpriyen
Posts: 17
Registered: ‎04-16-2013
Message 1 of 10 (387 Views)

/dev/async

[ Edited ]

what is the use of /dev/async file.

 

 

P.S. This thread has been moved from HP-UX > System Administration to databases - Hp forum moderator

 

 

Please use plain text.
Frequent Advisor
rveri-admin
Posts: 50
Registered: ‎06-25-2012
Message 2 of 10 (379 Views)

Re: /dev/async

It is used for oracle database for async communication.
Please use plain text.
Frequent Advisor
rveri-admin
Posts: 50
Registered: ‎06-25-2012
Message 3 of 10 (378 Views)

Re: /dev/async

 

More:

 

C.4 Asynchronous Input-Output

The asynchronous Input-Output pseudo-driver on HP-UX enables Oracle Database to perform Input-Output to raw disk partitions using an asynchronous method, resulting in less Input-Output overhead and higher throughput. You can use the asynchronous Input-Output pseudo-driver on both HP-UX servers and workstations.

This section contains the following topics:

Please use plain text.
Occasional Advisor
Deepanpriyen
Posts: 17
Registered: ‎04-16-2013
Message 4 of 10 (374 Views)

Re: /dev/async

>>The asynchronous Input-Output pseudo-driver on HP-UX enables Oracle Database to perform Input-Output to raw disk

 

but we are not using any raw file system.

 

but today we got infomation from DBA that trace files are getting generated abnormally.

 

oracle version 9.2 and hp -ux version  : v2.

 

1.what could be the reason for trace file getting generated in more numbers ?

 

2.if i am not using any raw disk,why this file is existing ? can I delete it ?

Please use plain text.
Honored Contributor
Patrick Wallek
Posts: 13,736
Registered: ‎06-21-2000
Message 5 of 10 (371 Views)

Re: /dev/async

1) Has the DBA looked at the trace files and the alert log?  It is the DBAs responsibility to figure this out initially.  It IS the database that is generating these after all.

 

2) There is no harm in this file existing on the system even if it is not in use.  I would not recommend deleting it.

Please use plain text.
Occasional Advisor
Deepanpriyen
Posts: 17
Registered: ‎04-16-2013
Message 6 of 10 (366 Views)

Re: /dev/async

DBA has told us that they get below information in Ioctl ASYNC_CONFIG error, errno = 1

 

and

 

1.they are asking us to change the permission for /dev/async to 660.

 

currently permission is 666 and owner is bin:bin.

 

2.i have tested in test servers and changed /dev/async permission to 660 and after running insf -e ,permission get changed to 666.

 

what is the reason after updating with  insf -e,permission for /dev/async is getting changed automatically.

 

 

 

Please use plain text.
Honored Contributor
Patrick Wallek
Posts: 13,736
Registered: ‎06-21-2000
Message 7 of 10 (361 Views)

Re: /dev/async

The changing of permissions of device files is a function of 'insf -e'.   Any time you run it, it will put the permissions of all device files back to their default values.

 

I do not recommend running 'insf -e'.  If you need to add new device files you should use one of the other options so you affect only specific device files and not everything.

Please use plain text.
Honored Contributor
Duncan Edmonstone
Posts: 5,678
Registered: ‎08-05-2000
Message 8 of 10 (343 Views)

Re: /dev/async

One would assume that your DBAs have been looking at Oracle support articles 1380006.1 or 302801.1 - but not very well...

 

setting /dev/async to 660 is just the first part of the task, they also have to:

 

(1) # /usr/sbin/setprivgrp dba MLOCK
(2) # vi /etc/privgroup
-- This should contain dba MLOCK RTSCHED RTPRIO
(3) # cat /etc/privgroup
dba MLOCK RTSCHED RTPRIO

 

And presumably restart the database...

 

What the Oracle article doesn't say is that actually the message is harmless, and also /dev/async remaining at 666 doesn't cause any problems either


HTH

Duncan
Please use plain text.
Honored Contributor
Duncan Edmonstone
Posts: 5,678
Registered: ‎08-05-2000
Message 9 of 10 (341 Views)

Re: /dev/async

... also that actually this is an Oracle bug in 9.2 where Oracle ignores the parameters:

 

disk_asynch_io=FALSE

filesystemio_options=none

 

in the init.ora file - this is fixed in 10.2 - you'll have to ask Oracle why they never back-ported the fix to 9.2...


HTH

Duncan
Please use plain text.
Trusted Contributor
Dave Chamberlin
Posts: 558
Registered: ‎04-10-2000
Message 10 of 10 (292 Views)

Re: /dev/async

I have seen this same problem with running 11G database on HPUX. Did all the steps indicated to fix the issue, but trace files are still generated continually with the same errors. Though they may be harmless - there is some overhead in generating 50000 trace files a day.

Please use plain text.
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation