stderr1 to 4

[Former Member]

Hi,

I would like to know if it is ok to delete or reset the stderr1 to 4 on this directory /usr/sap/SID/DVEBMGS00/work? If yes, send me the step by step procedure on how to do it and the use of those directories.

Another thing would be regarding the alertSID.log, Is it possible to delete or reset the content of this file? If yes send me the proceudre on how to do it.

Regards,

[Former Member]

Hi,

STDerr files are the log of error happen during system run state, they are not responsible for ri=unning of system but are usefull to analyse the issue getting triggered.

If you feel you wanted to empty(dont delete the files), move .old of them at OS level and create a empty file. But this activity has to be done only when system was in non-operative mode.

Alternatively you can reset the Log through SM50.

In general no body delete or moves the files since the amount of space occupied by them was considerbly less.

regards...

[Former Member]

Hi,

Just in addition to above, For deleting or to reset alert log file please refer SAP note "786032 - How to trim Oracle Alert log file"

Thanks,

Sushil

[Former Member]

Hi,

The files stderr0, stderr1, ... are, as the name indicates, the file to which the processes of the active SAP instance redirect their standard error. For standard output, the corrsepondilg files are stdout0, stdout1, ...

You can find the name of the process owning the file by looking at its first few lines, e.g. "dw.sap..." = dispatcher, "ms.sap..." = message server

As a rule you should leave these files in peace when the instance is running. Normally they do not grow very large (most error information is written to the dev_* traces) but if they do you should first of all check the errors that are reported and act on them.

When the instance is stopped it is perfectly OK to delete these files; they will be recreated at the next start.

As to alertSID.log, you should not delete or empty the file because it contains valuable information about the behaviour of the database. However, as the file grows continuously it tends to become unwieldy over time and the oldest information (I'vs seen alert logs with more than 6 years of data in them) will have lost most of its usefulness. A good approach is to periodically rename the alert log, e.g. on the 1st day of every quarter you could rename alert_SID.log to alert_SID_2009Q1.log, and so on. Oracle will automatically create a new alert_SID.log, no stop/restart of the database is needed.

What you should also do is to periodically clean up the old traces (*.trc) in the directories /oracle/SID/saptrace/background and /oracle/SID/saptrace/usertrace.

Hope this helps,

Mark

[Former Member]

Hi,

Another thing would be regarding the alertSID.log, Is it possible to delete or reset the content of this file? If yes send me the proceudre on how to do it.

Yes, alertSID.log can be deleted or renamed for further analysis. It will be created by oracle if it is not there. Please remember this file contains important log information like block corruption. Also this file must be checked before and after recovery.

Hope this helps.

Manoj

[Former Member]

HI,

Please do not delete the files in the directory /usr/sap/SID/DVEBMGS00/work.

You can reset them through SM50 and click on Trace and then in the drop down list you will Reset

In Reset you will find 2 options

1.Reset of work process files

2.Resetting of all files.

Select any one of those.

For the first it won't delete the data.Just it will rename the files as .old extension.

It will give you the required space.

If you do the second time it will delete .old extension files.Please monitor for few days and if the data is not required then you can do the second time.

Regards

Sreedhar Reddy

[Former Member]

Hello James,

You should not look to delete alertSID.log file.all the errros that come are logged in this file

what you can do is delete the old logs from this file so that the size gets controlled

Rohit