Eclipse: "Update SVN cache" hangs and locks up - java

Every time I start eclipse, the program begins doing this "Update SVN cache" thing but it will sit at 0% forever. I cannot perform any operations (such as checking out my projects, building them, or even running them) until this operation is completed (which never happens). Also, whenever I try to type anything in the editor, the whole program freezes and I have to kill the process.
I have been searching google for the answer to this problem for days and have yet to come up with an answer. Has anyone else had a similar problem and found a solution?
I'd like to add that I've tried re installing eclipse, all its plugins, and the jdk from scratch. Nothing seems to be working.

I faced the same issue and I tried to get out of this by disable most of performance setting for SVN in Eclipse:
Windows > Preferences > Team > SVN > Performance
OR (for latest versions): Windows > Preferences > Version Control (Team) > SVN > Performance
Disable: Computing deep outgoing state, Cache, persistent SSH

I just ran into this issue and was able to rescue it. This was with Zend Studio 10.5 which sits on Juno. I have about five projects in my workspace, one of which was open. I couldn't close the project because it was waiting for "Update SVN Cache" to complete.
With Eclipse closed, I went to my open project and via the command line ran "svn cleanup."
In workspace/.metadata/.plugins/org.eclipse.team.svn.core, I had a bunch of temp directories. I created a backup tarball, and then blew them all away.
That didn't fix anything. Finally I tried this:
With Eclipse closed, I went to my project directory and renamed .project to project.xml.
Reopened Eclipse, projects were closed. No SVN Update messages.
Restarted Eclipse.
Opened the project, but Eclipse balked at the missing .project file.
Closed Eclipse.
Went to my project directory and renamed project.xml back to .project.
Restarted Eclipse.
Opened the project. Smooth sailing. Was just able to commit a change without incident.
So far my workspace/.metadata/.plugins/org.eclipse.team.svn.core directory is still empty.
I don't know if the first two things I tried helped out at all, or if just renaming the .project file to force-close the project was all it took. Next time it happens (and there will be a next time) I'll try just force-closing the project and report back.

In my case I realized that a tortoise SVN explorer related windows was somewhere open. It probably locked the cache and eclipse was waiting for the unlock.
Not a direct answer to the question, but this might help for somebody making a similar mistake.

You're not the only one (see this bug report or this forum thread) but it's probably not a bug in Eclipse itself. Next steps:
Get a thread dump to see whether this is a deadlock or a thread is waiting for something that never happens (in the bug report, it hangs in System.loadLibrary()). You can use jconsole for this, it comes with the SDK.
Check all open projects in your workspace (that use SVN) with another SVN tool (command line svn or TurtoiseSVN if you're on windows) to make sure the data structures aren't corrupt.
Get the latest version of Eclipse and/or the SVN plugin
Try a different connector. Some people fare better with the JNI solution javahl, others with the pure-Java SVNKit.

Windows > Preferences > Team > SVN > Performance
Disable: Computing deep outgoing state, Cache, persistent SSH
Go to Preference -> General
Enable : Always run in background
Enable : Show heap status
Workbench Save interval (in minutes):9999
This will show your memory usage in eclipse.
Then edit your eclipse.ini file and change Xms and Xmx values to these:
-launcher.XXMaxPermSize 512m
-Xms1024m
-Xmx1024m

I faced this problem when having the same project in the workspace twice. Make sure you have only one copy of it in the workspace, maybe it helps.

It could be eclipse memory issue. I had similar problem until I did these steps:
Go to Preference -> General and put these values:
This will show your memory usage in eclipse.
Then edit your eclipse.ini file and change Xms and Xmx values to these:
-Xms1024m
-Xmx1024m
Look at the memory status. Hope it will help.

I was also facing similar issue and I got it resolved this by unchecking "Compute deep outgoing state for folders" under Windows->Preferences->Team->SVN->Label Decorations

This topic helped me for solving the problem about my eclipse svn update block.
I am using eclipse mars where this case came.
First I have deleted two projects which I don't needed - to reduce the project number in the workspace.
Changing parameter in eclipse.ini has not helped too.
I had the -clean argument on my eclipse set so I have deleted that too in hope it will work better. It helped a bit so the eclispe was running much longer but not much more. rockfarkas hint here.
What really helped at the end was to mark very fast all projects CtrlA from the project view and close them via the context menu while the SVN cache was still running. The half of them did that the others couldn't but after that I could see that the SVN cache was changing its state and so I could work again. The next step was to close the rest of the projects and open one by one. To this idea I came reading the post of user3096856.

I solved doing this:
-Select the project (or all projects) on your workspace (in eclipse), right click, select from the menu Team > Cleanup

I found an index error in eclipse which blocking the svn. Another project resources and svn repo were in .metadata\.plugins\org.eclipse.core.resources\.projects\ProjectName\4.tree
and other resources.
svn cleanup with the previous answer with 'Team > SVN > Performance' doesn't solve my problem.
'eclipse.exe -clean' will make eclipse unresponsible with 100% cpu.
The only solution was after exit eclipse the manual deleting of org.eclipse.core.resources\.projects\* files.

In my case, for 30 odd minutes it showed just 0 percent completed. I patiently waited since I tried restarting eclipse/machine still was getting the same thing. After 30 minutes, it continued with my svn update operation and completed it successfully.
Patience at times, helps :)

If there is duplicate project from SVN.
Force close eclipse.
Delete the duplicate project from filespace. (Eg: C:\Users\XXX\Workspace\DuplicateProject)
Restarting the eclipse
fixed.

I used Subversive - SVN Team Provider 3.0.0 with SVNKit 1.8.10 as SVN Connctor.
I updated the SVN Connctor to the latest version 1.8.11,and fixed it.

Do you use Maven SCM connector ? Try uninstalling it if you dont need it.
SCM connector is required for Maven SCM checkout, I do not use Maven SCM project checkout, instead I use SubEclipse checkout and convert project to maven using import existing project or convert to Maven project option from the context menu in Eclipse.

Related

Why is IntelliJ IDEA hanging on "Indexing"?

Want to improve this post? Provide detailed answers to this question, including citations and an explanation of why your answer is correct. Answers without enough detail may be edited or deleted.
I'm running on Arch Linux, on an i7-5930k 6 core CPU and 64GB of DDR4 RAM, and I'm using IntelliJ IDEA 14.
IDEA was working just fine for me several days ago, but one day, suddenly, it began hanging after opening a project, during the "Indexing" stage. I did not update IDEA and nothing changed about my projects. The IDE's UI hangs after it opens the project, with just a tiny little sliver of the progress bar for "Indexing" complete. Every 5-10 minutes or so it unfreezes and the progress bar crawls forward a little bit, before the IDE freezes again for another few minutes. This happens repeatedly for anywhere between 15 minutes and an hour, until it is finally finished indexing, at which point it hangs for another 5-10 minutes doing nothing, before it finally unlocks and allows me to develop.
While this is happening, my system is fairly unresponsive - Firefox tabs take a long time to switch, and scrolling in them is laggy. Opening a new terminal window takes a long time. Switching windows in general takes awhile. In htop, one of my CPU cores is loaded at 100% while the rest have a normal load, and about 6GB of RAM is used (fairly normal load when this system is idle.)
Things I have tried that have not helped:
Delete caches folder
Delete entire ~/.IntelliJIDEA14 folder
Reinstall IntelliJ package
Download IntelliJ manually from JetBrains' site and run it from my Downloads folder (to see if there was something wrong with the Arch AUR package)
Configure IntelliJ to use my system JVM and Maven for importing instead of its embedded tools
Opening multiple different projects (not just the one I initially experienced the issue on.)
This issue is really hurting my workflow, if anybody has any solution to this I would be very greatful.
Try Invalidating the cache and restarting IntelliJ.
In the File menu, select Invalidate Caches / Restart... and then click the Invalidate and Restart button.
I've finally figured it out. The solution was... Rather odd. TL;DR: Run it under strace. Read on for a more detailed explanation.
I came upon it when I decided to run IntelliJ under strace to see what files it was opening to determine whether or not it was a filesystem bottleneck.
This gave me some very strange results: strace was spewing out a near-constant stream of segfaults. Not only that, but IntelliJ was running just fine, not taking forever to index.
After consulting with a friend, I learned that on Arch Linux, systemd logs a dump of a process's memory every time a segfault occurs, except when a debugger is attached. strace is considered a debugger. Arch was thrashing my disks when it kept logging memory dumps due to all the segfaults, hence why the indexing was taking so long, because it was fighting for disk I/O.
My solution for now is to simply run IntelliJ under strace. I will, however, be looking into the issue further, as I don't think java should be segfaulting that much.
edit Intellij[VERSION]/bin/idea.properties,
set idea.max.intellisense.filesize=50
update:
Intellij will skip index files that size larger than 50kb.try this if you have many libraries or many large files(too many characters one line or too many lines)
I had this issue as well with version 2016.2 on Mac OS X. I had to do a force quit to end the application, then I deleted the .idea folder. The next time I launched IntelliJ everything worked fine, it had no problem indexing the project.
I was stuck with a similar issue with the latest IntelliJ Idea 2019.3, so maybe it'll help. For me the issue was with one of the plugins, uninstalling/reinstalling, cleaning caches didn't help. My steps were:
Kill intellij, start it over
When it's starting and about to load a project, be quick and cancel opening the project. This way you'll end up with a small window with a list of previously opened projects and a few menu items.
Open menu > plugins, disable them all
Restart intellij idea, open any project.
If step 4 above succeeds (which happened to me), one by one try enabling the plugins to see which one is causing error. For me it was Kubernetes plugin from JetBrains.
Select Help -> Debug Log Settings...
Add the following line (note the leading # symbol)
#com.intellij.util.indexing:trace
Relaunch the IDE (don't need to invalidate cache as that will cause it to start from scratch, whereas restarting from the point of failure, for me anyway, reported the problem file as soon as I restarted):
Scheduling indexing of file://C:/dev/tools/ruby/lib/ruby/2.2.0/x64-mingw32/win32ole.so by request of index Stubs
Our project doesn't use win32ole so I moved the file to a safe location and restarted my IDE... Bingo, problem gone, indexing finally completed after almost 1 year of effectively using intellij as a slightly-smarter-than-notepad ruby editor.
In my case I found out that Intellij is actually trying to index a 50GB directory with logs that was under the project's root. Make sure that if you have such a directory, it is marked as "Excluded" in the IDE.
You can see which file the IDE is indexing currently in the Indexing Status window (access by clicking on the indexing message in the toolbar). You may need to enlarge this window to see the full path of the file currently being indexed.
In PhpStorm, what solved this for me was excluding folders I didn't need to be indexed from the indexing (specifically the vendor folder, a caches folder, and a few asset folders that contained thousands of images). Instantly it began making progress and completed.
To do this:
in the project directory list, right click the folder you want to exclude
Mark Directory As
Excluded
seems that there might be several reasons for getting into this "indexing" hell. I spent few hours trying to fix it using the ideas above.
At the end of the day, with some profiling work, I found that the bad guy in my case was the csv plugin:
https://plugins.jetbrains.com/plugin/10037-csv-plugin
I had few (not so large) CSV files serving as input, and although I marked them as not to be indexed, the plugin kept trying to index them.
Once I removed the plugin everything worked fine.
Disabling unused plugins will improve the indexing. In my case I have disabled Kotlin plugin from File -> Settings->Plugins
if you check the intellij logs (can be found under C:\Users<User Name>.IntelliJIdea2019.1\system\log ) you will get a pointer what is failing. I was getting an error in Kotlin. After disabling the plugin and restarting Intellij fixed my issue
java.lang.RuntimeException: java.io.EOFException
at com.intellij.util.ExceptionUtilRt.rethrow(ExceptionUtilRt.java:31)
at com.intellij.util.ExceptionUtil.rethrow(ExceptionUtil.java:120)
at com.intellij.openapi.vfs.newvfs.persistent.FSRecords$DbConnection.handleError(FSRecords.java:516)
at com.intellij.openapi.vfs.newvfs.persistent.FSRecords$DbConnection.access$000(FSRecords.java:153)
at com.intellij.openapi.vfs.newvfs.persistent.FSRecords.writeAndHandleErrors(FSRecords.java:965)
at com.intellij.openapi.vfs.newvfs.persistent.FSRecords.access$300(FSRecords.java:47)
at com.intellij.openapi.vfs.newvfs.persistent.FSRecords$AttributeOutputStream.close(FSRecords.java:1629)
at kotlin.io.CloseableKt.closeFinally(Closeable.kt:53)
at org.jetbrains.kotlin.idea.caches.FileAttributeServiceImpl.write(FileAttributeServiceImpl.kt:64)
at org.jetbrains.kotlin.idea.caches.FileAttributeServiceImpl.writeBooleanAttribute(FileAttributeServiceImpl.kt:48)
at org.jetbrains.kotlin.idea.caches.IDEKotlinBinaryClassCache.getKotlinBinaryClass(IDEKotlinBinaryClassCache.kt:67)
at org.jetbrains.kotlin.idea.caches.IDEKotlinBinaryClassCache.getKotlinBinaryClassHeaderData(IDEKotlinBinaryClassCache.kt:8
I had a similar problem with 2019.1.4. However, mine would change directories and sometimes, eventually, finish. If it did finish, it was somewhere in the 8-10 minute range.
I was all over SO, and even JetBrains' forums. I excluded directories via Project Structure | Modules. I used Invalidate Caches And Restart on multiple occasions. I tried having only 1 project open to let it finish. I installed and tried 2019.2.4 and 2019.3.3 (the latter would crash for other reasons). And best of all, it only seemed to happen on one project!
What ultimately led me to an answer was Help > Activity Monitor... where I found psi.impl.cache.impl.todo was running at nearly 100% CPU and not showing signs of stopping.
It turned out, I had a TODO filter setup with a poorly-defined RegEx. It was something like \b.*wip\b.*; the idea was to find all of our "WIP" values. Well, the leading .* was a huge mistake and one that didn't occur to me until losing many hours blaming a plugin upgrade. I believe the reason this was a bad filter was because the project it was hanging on is in ExtJS, which is JavaScript, which means things are in triplicate with the app.js file and whatnot...
Update
In IntelliJ 2020.x and 2021.x, the option has moved to Help > Diagnostic Tools > Activity Monitor.
Had the same issue in the past on some Scala project. I have installed IDEA 16 EAP (https://confluence.jetbrains.com/display/IDEADEV/IDEA+16+EAP) and the problem has gone.
I ran into this problem today on a Mac. It would hang before I was able to get to the menu and invalidate the cache.
I deleted the cache from command line using the following command and it worked for me.
rm -rf ~/Library/Caches/JetBrains/IntelliJIdea*
After that it started up with no issues.
I had the same problem with IntelliJ 2017.3.2. When I clicked on the indexing progress bar I noticed it was hung on a directory within my build directory. When I did a gradlew clean which removed that directory then the indexing was able to proceed.
I have encountered this problem, and resolved it:
remove idea
delete all files and dirs which name regex 'jetbrain' and 'IntelliJ' in my computer(Mac mini)
then install idea
I also try just delete idea cache files, it do not work.
I was using the Elm plugin and installed the elm-bounded-nats package which included a semi-large source file. IntelliJ kept hanging on this file, but did not always report this correctly in the indexing popover dialog (perhaps due to threading). When I exluded this specific file in Settings -> File Types ("Nats.elm") indexing managed to complete successfully. Now the editor renders errors for this package but the compilation process still works.
Invalidate Cache and Restart did not work for me on IntelliJ 2021.3 Ultimate.
I was curious to see if the new Repair IDE feature on 2021.3 Ultimate version works on this issue on my mac.
On startup, IntelliJ froze on indexing
Force quit and relaunched IntelliJ
On Startup, manually "paused" the indexing from the status bar at the bottom
Invalidated cache and restarted
Still froze on indexing
Forced quit and relaunched IntelliJ
Repeated Step #2 on startup
From File Menu, executed Repair IDE, and went through repair steps
FIXED; no longer froze on indexing
Quit IntelliJ and Relaunch
No Index issues
I was also facing freezing issue with intellij 2021.3. Earlier I was using intellij 2021.1 version and that was working fine but since I upgraded intellij version to 2021.3 it started freezing on indexing of files.
Someone in this thread suggested to repair ide using Repair IDE feature. But that didn't worked for me.
So, I went through the thread dumps available in intellij logs folder. After analyzing logs, I found out that calls are blocking on Package Search plugin. So, I disabled that plugin in settings -> plugin. To do that I had to pause indexing at start for time being. After disabling this plugin, it is working fine for me.
This is a known issue and I know disabling it is not a correct fix for it. But I will use it till intellij will release an official fix for this issue.
There might be another project opened parallelly in new window which is being indexed.
It looks like the problem can come from many different sources as other answers point out.
In my case, it was the Subversion plugin that had difficulties to communicate with the server and make the IDE to hang on indexing.
I was able to resolve this problem by removing all of my "target" folders from my project.

How to fix corrupted Eclipse Indigo JEE?

I installed following plugins in eclipse indigo in following order to start spring development
Spring Tools Suite
M2E (Maven)
after these two installations, it was giving error
jira connector not installed, so I installed the following plugin.
Atlassian Jira Connector
After installing Jira Connector, Eclipse Started showing the following error :
Uninstalled Jira Connector. Still Showing these Problems.
Any help would be appreciated.
Eclipse has major structural problems with uninstalls, which aren't really fixed even the latest luna. But don't worry, there are a lot of workarounds.
What you can do now and in similar situations:
1.
If you can even uninstall something, you won't get back its previous state before the install.
Because of it I use Eclipse normally with a trick: I store my main Eclipse install directory in a git repository, and so I can always switch back with a single command. But it is only a trick.
2.
There is a big chance, that only your workspace directory is damaged, and not your Eclipse. In this case you can solve this problem by reinitializing your workspace: make a backup, delete everything, recreate your workspace directory and finally import again your projects.
(For similar reasons it is also an useful trick to save your workspace metadata in a git repository as well.)
3.
In Eclipse, the menu items are created by modules. If an eclipse module is installed, it creates the changes in its internal configurations which create the menu items.
After a restart, Eclipse tries to restore your gui, and thus re-open its panels. But if a module uninstall is also happened, then its panels aren't restorable, resulting exactly your problem.
So, simply close the bad panels and try to reopen them. Sometimes it also works.
In short: recreate your workspace, it will probably help. And next time, use Eclipse with some good and frequent backup (I suggest git).

Eclipse java build slower after moving project to GIT

I'm new to GIT and recently moved over one of my eclipse java projects.
I noticed a significant delay in the auto build of my project after moving it to GIT. Seems to be some GIT update that happens on every save.
"Updating GIT status for repository GIT"
I'd appreciate any insights.
Eclipse version: Luna RC3 Release (4.4.0RC3)
I'm not sure if this is the culprit but almost every time in the last couple of Eclipse releases I have had issues with the Remote System Explorer which often for some reason gets called when you have source control plugins or aspectj on a project.
For a git project on Luna I turned it off by following:
"Remote System Explorer Operation" causing freeze for couple of seconds
Yes the answer is for a different problem but again I have found the Remote System Explorer to be the culprit (or maybe indirectly I don't know for sure but I do know when I turn it off things get faster). There is also some git specific answers to that referenced SO question but I found I had to turn RSE completely off.

Eclipse Workspace Corrupted due to improper shutdown

I'm working with Eclipse juno on windows7-64bit OS. Due to sudden power problem my system got shutdown, when i restart the Eclipse next start, Eclipse fails to load the workspace. It freezes on startup, or the workspace does not show up. It seems the workspace gets corrupted every time Eclipse is not shut down properly.
One way to fix the startup is
rm -rf ~/workspace/.metadata
After doing this, of course, settings are gone, projects have to be reimported etc. - It's really a pain :( and I don't understand why it has to be this way. Other applications seem to be able to keep their data intact, even if they are killed.
Can you suggest ways to remedy this problem? Are there ways of recovering a corrupted workspace including the settings?
thanks in advance
You can edit your shortcut to eclipse and add behind the line -clean as runtime parameter. It should clean your workspace and it might fix your corruption.
http://help.eclipse.org/indigo/index.jsp?topic=%2Forg.eclipse.platform.doc.isv%2Freference%2Fmisc%2Fruntime-options.html
So your shortcut would be something like:
"C:\Program Files(x86)\Eclipse\eclipse.exe" -clean
EDIT: Make sure once you've used this to remove it again after, you don't want to clean your eclipse every time you start it.
from Vissol.ca
create a new workspace
import new (android) project(s) from existing source(s)
good to go...
then you can backup original workspace , delete it and recreate it if desired.

Eclipse 3.6 freeze on autocomplete/quick fix

I have a problem that seems to come up both with autocomplete and quick fix. Sometimes when I use cmd+1 for quick fix dialog, Eclipse freezes and however long I wait (30 min at least), nothing happens. When inspected in activity monitor, it seems to exhibit little to no processor activity.
I have a Mac with OSX 10.6, and Helios with no weird plugins. I program in java, so I use very standard features.
Is this a known bug? I have tried to google a lot to find useful info. I am not very good at reading bug reports though, and the concept of feature freeze kind of taints my search results.
Thanks for any help.
As for Eclipse you can not know what is happening. The whole platform is a set of plugins, and the requirements (for example responsiveness) are not strict for them. So a third party or even a bundled plugin can cause such a failure. So most of times you cannot have a clue about what is wrong. Some stuff you can do:
while hanging unplug your network connection. If you are behind a proxy for example, a plugin can wait for a long time on network io
you can check with for example resource monitor, which file eclipse opened. Some of the opened file handlers can have relation with the problem
edit .classpath file in project where code completion freezes, and it will "reset" your project stuff in eclipse
you can browse .metadata folder in eclipse workspace and you can guess which folder to remove temporaly. After removing, try if error is still present
create a new workspace and import projects into it
if new workspace do not work, then the eror is in configuration folder in eclipse root, and you can play the same thing like in step 4
use a brand new eclipse
+1. Maybe some new plugin is responsible. In eclipse Help/about/installation details/Installation history tab you can revert to some older set of plugins.
Hope it will help.
Seems that you encounter known and already fixed bug.
Issue fixed after deleting all the files in the below directory
configuration\org.eclipse.osgi\.manager
.fileTableLock
.fileTable.1
.fileTable.2
There will be fileTable lock files, delete all those files

Categories

Resources