[ https://issues.apache.org/jira/browse/TIKA-1796?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15008971#comment-15008971 ]
Nick Burch commented on TIKA-1796: ---------------------------------- Firstly, please don't post to the dev list, ignore the response then open a jira for the very same thing! As explained in response to your post to the dev list, Apache Tika 0.9 is very old now. There have been lots of fixes since then, including in areas like this Please upgrade to the most recent version (1.11) and retry > Issues with tika jar and Microsoft documents like doc.,ppt, xls etc > ------------------------------------------------------------------- > > Key: TIKA-1796 > URL: https://issues.apache.org/jira/browse/TIKA-1796 > Project: Tika > Issue Type: Bug > Affects Versions: 0.9 > Environment: UNIX server > Reporter: Femi > > We have had a problem with tika-app-0.9.jar when it comes to using Microsoft > documents (we do not have issues with PDFs and images). It creates tika files > which are held by our weblogic java process. > For example, if one runs the command :- lsof -p 27305|grep deleted > java 27305 oracle 330r REG 253,1 295674 68 > /tmp/apache-tika-5125182301796025972.tmp (deleted) > java 27305 oracle 334r REG 253,1 272896 69 > /tmp/apache-tika-8997882426533237375.tmp (deleted) > java 27305 oracle 335r REG 253,1 295674 78 > /tmp/apache-tika-5232377327199509251.tmp (deleted) > java 27305 oracle 336r REG 253,1 45327 43 > /tmp/apache-tika-6884061409786039638.tmp (deleted) > java 27305 oracle 339r REG 253,1 272895 41 > /tmp/apache-tika-6752501215118342524.tmp (deleted) > java 27305 oracle 340r REG 253,1 272895 41 > /tmp/apache-tika-6752501215118342524.tmp (deleted) > java 27305 oracle 341r REG 253,1 45327 75 > /tmp/apache-tika-7548218713808428132.tmp (deleted) > The above is a long list of held tika files from Microsoft docs in deleted > state but they are still handled by the weblogic process. > The only way we can get these tika files closed or released is by restarting > the weblogic server. > This cost us money as we had to stop the server to get rid of the tika files > filling up our tmp folder. > We have had this issue for almost 3 years now. I have been researching on the > web to see if there are solutions out there in an upgraded tika-jar but it > seems there are none. > I was thinking it will be resolved in an upgraded jar file but it seems that > is not the case. > Please is there any solution to this issue? > Regards, > Femi Balogun, > Application Support Engineer, -- This message was sent by Atlassian JIRA (v6.3.4#6332)