Discussion:
Temporary files not automatically deleted
(too old to reply)
Keith
2023-11-02 05:47:14 UTC
Permalink
DSAP shows a large amount of temp files on a disk. On closer investigation I found out that there are lots of temp files but are not opened by any processes. This should not happened as closing a temp file should have it automatically deleted. Attempt to manually purge them returns an error 12. Has anyone encountered this and how to fix this?
JShepherd
2023-11-02 13:59:21 UTC
Permalink
DSAP shows a large amount of temp files on a disk. On closer investigation =
I found out that there are lots of temp files but are not opened by any pro=
cesses. This should not happened as closing a temp file should have it auto=
matically deleted. Attempt to manually purge them returns an error 12. Has =
anyone encountered this and how to fix this?
A bigger problem, on systems with a lot of SQL/MX create/drop happening,
is the mass of .ddl files in /usr/tandem/sqlmx/ddl .

A single QA test cycle on MX here can create/drop a thousand MX tables.

The first time /usr became full it took a while to find the culprit,
which was was over 300,000 .ddl files in /usr/tandem/sqlmx/ddl

Now, we have a program we run occasionally to clean them out

From this morning on one system
mxscan has deleted 14068 .ddl files
Randall
2023-11-02 22:03:54 UTC
Permalink
Post by JShepherd
DSAP shows a large amount of temp files on a disk. On closer investigation =
I found out that there are lots of temp files but are not opened by any pro=
cesses. This should not happened as closing a temp file should have it auto=
matically deleted. Attempt to manually purge them returns an error 12. Has =
anyone encountered this and how to fix this?
A bigger problem, on systems with a lot of SQL/MX create/drop happening,
is the mass of .ddl files in /usr/tandem/sqlmx/ddl .
A single QA test cycle on MX here can create/drop a thousand MX tables.
The first time /usr became full it took a while to find the culprit,
which was was over 300,000 .ddl files in /usr/tandem/sqlmx/ddl
Now, we have a program we run occasionally to clean them out
From this morning on one system
mxscan has deleted 14068 .ddl files
You might want to look at how swap space is configured on your system. Some of these files (particularly the ones you cannot purge) might be associated with those.
Keith
2023-11-06 00:59:13 UTC
Permalink
Post by JShepherd
DSAP shows a large amount of temp files on a disk. On closer investigation =
I found out that there are lots of temp files but are not opened by any pro=
cesses. This should not happened as closing a temp file should have it auto=
matically deleted. Attempt to manually purge them returns an error 12. Has =
anyone encountered this and how to fix this?
A bigger problem, on systems with a lot of SQL/MX create/drop happening,
is the mass of .ddl files in /usr/tandem/sqlmx/ddl .
A single QA test cycle on MX here can create/drop a thousand MX tables.
The first time /usr became full it took a while to find the culprit,
which was was over 300,000 .ddl files in /usr/tandem/sqlmx/ddl
Now, we have a program we run occasionally to clean them out
From this morning on one system
mxscan has deleted 14068 .ddl files
Raised this with HPE. The issue was that a permanent file with CLEARONPURGE was purged. DP2 converts this to a temp file but did not purge it. It's a DP2 issue unfortunately.
Loading...