Why would tmp directory under Tibco Data Virtualization (TDV) have certain files in it?
Article ID: KB0082513
Updated On:
| Products | Versions |
|---|---|
| TIBCO Data Virtualization | 7.0.6 |
Description
Why would tmp directory under Tibco Data Virtualization (TDV) have certain files in it?
Introduction
When using Tibco Data Virtualization (TDV), you might observe that the disk usage varies automatically under TDV <X.X.0>/tmp directory.
Introduction
When using Tibco Data Virtualization (TDV), you might observe that the disk usage varies automatically under TDV <X.X.0>/tmp directory.
Issue/Introduction
Why would tmp directory under Tibco Data Virtualization (TDV) have certain files in it?
Resolution
Files will be created in Tibco Data Virtualization (TDV) under TDV X.x.0/tmp for following cases:
Temporary files created will be cleaned up automatically in the following cases:
#1 cursor
#3 dataship
Following are the names of files/directories that will get created.
Cursor Temp Files:
hashcursors
distinctcursors
pipecursors
resetcursors
seekcursors
sortedcursor
dynamicmap
dynamicmapext
preparedmap
streamtransform/index
streamtransform/value
Cluster Files:
cluster/archive.car
— SEND RESULTS IN EMAIL STORED PROCEDURE —
OUTPUT_<>_TMPFILE.CSV FILE
OUTPUT_<>_TMPFILE.XML
— DATASHIP Temp Files---
dataship/*
— CACHE Files---
cache/csstatus/csstatus
cache/csstatus/meta
filecache
When a request is served from disk, the memory in use decreases and the query engine creates files under <Tibco Data Virtualization (TDV)>/tmp/ subfolders (for example: < Tibco Data Virtualization (TDV)>/tmp/hashcursors folders) . These folders contain the in-memory federated join swaps to disk
for that request. This will also happen if you use
{ OPTION FORCE_DISK }
or
{ OPTION FORCE_DISK2}
in the query.
These tmp tables are created when we "spill to disk" when the results of a query are too large to hold in memory and the results from 2 data sources are still being processed with a hash sort.
If the timestamp of the files in< Tibco Data Virtualization (TDV)>/tmp/ subfolders were from today, it's possible that the query was still running, especially if the files couldn't be deleted - TDV may still have had them open.
These files are supposed to be removed after the join is completed.
Please make sure no queries are running before you try to remove the files under the <Tibco Data Virtualization (TDV)>/tmp/ subfolders. Otherwise, you can stop the Tibco Data Virtualization (TDV) server and these files will be deleted automatically during a Tibco Data Virtualization (TDV) restart.
- Cursor failing over to disk
- CIS stored procedure
- Dataship feature
- File cache
- CIS cluster log
Temporary files created will be cleaned up automatically in the following cases:
#1 cursor
#3 dataship
Following are the names of files/directories that will get created.
Cursor Temp Files:
hashcursors
distinctcursors
pipecursors
resetcursors
seekcursors
sortedcursor
dynamicmap
dynamicmapext
preparedmap
streamtransform/index
streamtransform/value
Cluster Files:
cluster/archive.car
— SEND RESULTS IN EMAIL STORED PROCEDURE —
OUTPUT_<>_TMPFILE.CSV FILE
OUTPUT_<>_TMPFILE.XML
— DATASHIP Temp Files---
dataship/*
— CACHE Files---
cache/csstatus/csstatus
cache/csstatus/meta
filecache
When a request is served from disk, the memory in use decreases and the query engine creates files under <Tibco Data Virtualization (TDV)>/tmp/ subfolders (for example: < Tibco Data Virtualization (TDV)>/tmp/hashcursors folders) . These folders contain the in-memory federated join swaps to disk
for that request. This will also happen if you use
{ OPTION FORCE_DISK }
or
{ OPTION FORCE_DISK2}
in the query.
These tmp tables are created when we "spill to disk" when the results of a query are too large to hold in memory and the results from 2 data sources are still being processed with a hash sort.
If the timestamp of the files in< Tibco Data Virtualization (TDV)>/tmp/ subfolders were from today, it's possible that the query was still running, especially if the files couldn't be deleted - TDV may still have had them open.
These files are supposed to be removed after the join is completed.
Please make sure no queries are running before you try to remove the files under the <Tibco Data Virtualization (TDV)>/tmp/ subfolders. Otherwise, you can stop the Tibco Data Virtualization (TDV) server and these files will be deleted automatically during a Tibco Data Virtualization (TDV) restart.
Feedback
Yes
No
Powered by
