[
https://issues.apache.org/jira/browse/CXF-7396?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17099645#comment-17099645
]
Mats Aune commented on CXF-7396:
--------------------------------
[~coheigea] This problem is still valid, ref screenshot of open files below ().
We updated from 3.1.5 to 3.3.4 and faced same issue. Tried to update to latest
3.3.6 but also see same problem there. We have two production environments but
the problem is more present in the env where the cxf service is invoked more
frequently and with larger responses. For example this is an excerpt the
content of one of the cxf-tmp-xxx dirs:
{code:bash}
-rwxr-xr-x 1 USER1 staff 3.9M Apr 24 14:48 cos5477634023836053854tmp
-rwxr-xr-x 1 USER1 staff 1.6M Apr 24 14:42 cos3679655186586487675tmp
-rwxr-xr-x 1 USER1 staff 1.5M Apr 24 14:44 cos8675304298389496862tmp
-rwxr-xr-x 1 USER1 staff 1.4M Apr 24 14:33 cos118758002212206609tmp
-rwxr-xr-x 1 USER1 staff 1.3M Apr 24 14:32 cos2144891793092235945tmp
-rwxr-xr-x 1 USER1 staff 1.3M Apr 24 14:46 cos1780262235511988201tmp
-rwxr-xr-x 1 USER1 staff 1.3M Apr 24 14:38 cos7829498659380896458tmp
-rwxr-xr-x 1 USER1 staff 1.3M Apr 24 14:46 cos6504563011674464168tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:45 cos2152216273289241192tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:41 cos3738419480920900092tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:44 cos1658911959056771225tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:37 cos2909352661354947925tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:49 cos8169790396201753344tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:42 cos2349963537403371904tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:43 cos619314010609582924tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:38 cos8595265956697142864tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:35 cos2407701855954590120tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:44 cos3530647692316061540tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:33 cos284863373279157205tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:48 cos8271202711406949622tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:39 cos5319391463654487053tmp
-rwxr-xr-x 1 USER1 staff 1.2M Apr 24 14:48 cos1458994500916581464tmp
{code}
I am not sure if this might be related to LoggingInInterceptor as this class is
deprecated.
I have done two things: Removed use of default logger (LoggingInInterceptor) in
code, added use of custom implementation (in spring config) that limits logging
size. If this does not work I will have revert to the version of cxf we used
earlier. At a later stage I can try to update cxf version AND logging
mechanisms currently in use marked as deprecated. !Screenshot 2020-05-05 at
10.09.43.png!
A lot of other applications is using the same cxf version without problems, so
probably related to a particular configuration of cxf and related
(LoggerInTerceptor??) code.
> CachedOutputStream doesn't delete temp files
> --------------------------------------------
>
> Key: CXF-7396
> URL: https://issues.apache.org/jira/browse/CXF-7396
> Project: CXF
> Issue Type: Bug
> Components: Core
> Affects Versions: 3.1.8
> Reporter: Matthew Roth
> Priority: Minor
> Fix For: NeedMoreInfo
>
> Attachments: HTTP Request.jmx, Screenshot 2020-05-05 at 10.09.43.png,
> image-2018-05-29-13-58-03-109.png, image-2018-05-29-14-00-16-442.png,
> image-2018-05-29-14-00-54-215.png
>
>
> In the CachedOutputStream the method maybeDeleteTempFile doesn't always
> delete the temp file when it should.
>
> this.streamList.remove(stream)
> Doesn't remove the stream, occasionally the stream is not in the list causing
> the check of this.streamList().isEmpty() to fail. Also occurs when
> this.streamList() contains multiple streams.
> This seems occur when too many large requests are processed in a row.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
