HIPE known issues

help Welcome to the known issues page. Please take a look at the table of contents below. You can find known issues for each HIPE version, in addition to installation/uninstallation, startup, help, Science Archive and other issues common to different HIPE versions.

ALERT! This page is for known issues in the HIPE software. For known issues in data products see this page.

ALERT! These known issues happen in the latest versions of HIPE and its dependent software (for example, Java 6/7). Any issue that is resolved with an update to the operational version of HIPE or Java, is listed on this page.

Please note that some issues detected in previous versions may still be present in HIPE 11. Scroll down and check earlier versions if you do not see your particular issue listed for HIPE 11.

Installation and de-installation issues

Installer or start-up hang on Mac OSX and Java 1.6.0_51

Ticket number: HCSS-18288

    • Some users have found that the InstallAnywhere GUI hangs leaving all buttons frozen; this may happen either when installing, or when starting HIPE after an apparently successful installation. This appears to be due to a failed Apple java update and can, in principle, occur with any HIPE version. You can check for this issue following these steps:
    • How to detect a failed update. Run: /usr/libexec/java_home -v 1.6 -exec java -version in a terminal. If the output includes "M4508" you have the bad update.
    • How to fix: Users need to manually download and install the Java update.
    • Apple support notes that "This update corrects an issue with Java applications failing to draw or respond to user input."

The installer for Mac complains about write permissions. Why?

You need write permissions for the directory where you want to install HIPE. For example, you need Administrator privileges to install in /Applications. If you have installed HIPE successfully in the past, and are now experiencing write permission problems, it may be that some other application changed the permissions in the meantime. There have been cases of scientific applications such as SciSoft doing this. Once you revert write permissions to their correct values, HIPE installation should succeed.

The installer for Windows does not start and issues the error message Not a valid Win32 application. Why?

This problem has been encountered under Windows Vista because the installer had not been downloaded correctly, resulting in an empty file. The exact cause is not yet known. As a workaround, try downloading the installer in a different way (for instance, via command line FTP rather than a web browser) or from another computer.

Why does the uninstaller leave some icons and files/directories behind?

This incomplete uninstallation is a known issue and may be caused by a corrupted InstallAnywhere global registry. As a workaround, try renaming the InstallAnywhere registry file. See point 7 in the InstallAnywhere issues page for detailed information. If you add files (such as Jython scripts) into HIPE directories after installation, these will not be removed by the uninstaller.

Sometimes when you run HIPE on your laptop for the first time you may get following error message:

% ./runhipe Error occurred during initialization of VM

Could not reserve enough space for object heap

Could not create the Java virtual machine.

This error typically happens because you have either set or have as default a too large memory for the Java Virtual Machine and hence it cannot start. Edit the file called "installed.properties" under your hipe_v11_RC2 (or alternative directory, as appropriate) installation directory and edit the line

java.vm.memory.max=2872m

to have a smaller value than 1500m and smaller than the RAM memory of your machine.

For other issues affecting the installer, check this page.

Known issues in HIPE 12

SUSSEXtractor returns many more sources for SPIRE maps from version 12.66 onwards

Ticket number: HCSS-18988

Versions affected: All versions of HIPE 12 from build 66 (included)

Description: Running the SUSSEXtractor task on the same SPIRE maps using versions of HIPE from version 12.66 onwards returns 3-4 times the number of sources it returned previously with the same parameters (excepting the PRF grid size, which has increased internally). The fluxes of the detected sources are also different and generally higher.

Workaround: You can force the old results by imposing a 5x5 PRF grid size, but tests are ongoing to decide whether the current settings produce more reliable photometry. Example script:

myImage = SimpleImage("your SPIRE map")
myPrf = PrfGaussian([myImage.getWidth(),myImage.getHeight()],5.0)
sourceList = sourceExtractorSussextractor( \
    image = myImage,        # Image name \
    ...
    prf = myPrf,            # 5x5 PRF \
    ... 
    )

SmoothSpectrumTask and ResampleSpectrumTask do not propagate weights correctly

Ticket number: HCSS-15583

Versions affected: All versions of HIPE since 8.0

Description: If any of the tasks is applied to a spectrum that contains a column of errors instead of weights, the task converts the errors to weights, propagates the weights and repopulates the error column by converting the propagated weights back to errors. This gives an error column that has not been propagated correctly for the operation applied.

Workaround: Take into account that the output error column should be disregarded in the case of source data that lack a proper error column.

HIPE FITS Writer format could be incorrectly read by other software

Ticket number: HCSS-19248

Versions affected: HIPE 12.x

Description: HIPE makes use of OGIP, a non-standard convention to include long values in FITS fields. The OGIP convention writes an & at the end of a line (for example, describing a column name) that would be considered valid for all FITS compliant readers (it contains less than 68 characters) followed by the special keyword CONTINUE '' / & and completing the long value with a COMMENT Rest of value line. For example:

TTYPE1  = 'designation&'
CONTINUE  '' / &
COMMENT Original element type: char, Empty values appear as null
Many FITS readers ignore this convention, so the special character & is read and included as the final character of the value. Others (IDL) do the same but, additionally, consider & an invalid character and it is replaced by an underscore _.

Workaround: None at this time.

Old syntax using HIFI class PolarPair gives wrong results

Ticket number: none

Versions affected: HIPE 12.0

Description: PolarPair (the class) use is now discouraged for users and polarPair (the task, note capitalisation) has been modified in order to allow it to be used as a part of the interactive HIFI pipeline and to bring its syntax in-line with other tasks in the Spectral Toolbox. If the old syntax is used, e.g.,

pp = PolarPair(wbs_h, wbs_v)
av = pp.avg()

incorrect results are obtained.

This problem did not affect previous HIPE versions. A warning will be written to the console in HIPE 13 if the old syntax is used.

Workaround: Use the new syntax running the task:

av=polarPair(ds1=wbs_h, ds2=wbs_v)

Problem with doGridding in rotated maps with HIFI

Ticket number: HIFI-5716

Versions affected: HIPE 12.0

Description: for maps taken with a non-zero orientation on the sky, the HIFI pipeline will generate two cubes: one non-rotated cube projected onto the Equatorial frame (cubesContext) and one rotated by the scanning orientation angle (cubesContextRotated). In HIPE 12.0 a bug was introduced that generates both cubes in the rotated frame.

Workaround: In order to generate non-rotated cubes you can apply the doGridding task on the level 2 products with the option flyAngle=1e-5 (for example - it is just a matter of using a very small, non-strictly zero, angle). The problem will be solved in 12.1.

computeVelocityTask plots absorption profiles as emission

Ticket number: HCSS-19133

Versions affected: HIPE 12.

Description: The computeVelocityTask in the Cube Analysis Toolbox returns and plots a fittedLineCube. This plot displays absorption profiles as emission, even if isEmission=False. The data returned is correct, but this plot must be disregarded.

Workaround: Simply disregard the plot generated by the task and trust the data stored in the fittedLineCube.

Limitations of the task for updating scripts that use the old syntax to access return parameters

Ticket number: HCSS-19080

Versions affected: HIPE 12.

  • Task calls spanning multiple lines (with \ separation) are not detected. Some SPIRE scripts use multiple line task call scripts.
  • The dialog box should show all lines which are going to be changed, but only the first line is shown there.
  • Some non-standard task calls, like using two = signs to put the result into two variables at once, are not detected.

Old Java 6 entries in the dynamic library environment variable cause HIPE 12 to crash on OS X

Ticket number: HCSS-18891

Versions affected: HIPE 12.

Error message:

$ ./hcss.dp.pacs-12.0.1553/bin/hipe
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x000000010794f647, pid=49119, tid=6403
#
# JRE version: Java(TM) SE Runtime Environment (7.0_45-b18) (build 1.7.0_45-b18)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (24.45-b08 mixed mode bsd-amd64 compressed oops)
# Problematic frame:
# C  [libjvmlinkage.dylib+0x3647]  JVM_GetClassCPEntriesCount+0x17
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# An error report file with more information is saved as:
# /Users/USER/hcss/hcss.dp.pacs-12.0.2340/bin/hs_err_pid49119.log
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.sun.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#

Root cause: The cause is that the shared library path has references to the old Java 6 installation.

$ env | grep DYLD_LIBRARY_PATH
DYLD_LIBRARY_PATH=:/Applications/MATLAB/MATLAB_Compiler_Runtime/v716/runtime/maci64:
/Applications/MATLAB/MATLAB_Compiler_Runtime/v716/sys/os/maci64:
/Applications/MATLAB/MATLAB_Compiler_Runtime/v716/bin/maci64:
/System/Library/Frameworks/JavaVM.framework/JavaVM:
/System/Library/Frameworks/JavaVM.framework/Libraries
Workaround: The current workaround is to remove the entries that contain JavaVM from this environment variable. To do that, modify or create the .bash_profile configuration file of the terminal and add the environment variable in an export statement without the corresponding entries. That should persist the changes.

Known issues in HIPE 11

Reading observations stored as tarball archives that contain new data types does not work in HIPE 11

Ticket number: HCSS-19572

Versions affected: Potentially all versions older than HIPE 12.

Error message: While loading an observation from a tarball archive (.TAR file), HIPE returns the following error:

WARNING: Could not open /home/user/1342219630-herschel.ia.obs.ObservationContext-580157.xml: 
herschel.ia.task.TaskException: Error processing getObservation task: 
Class definition not found for urn urn:hsa:herschel.spire.ia.dataset.PhotApertureEfficiency:0: herschel.spire.ia.dataset.PhotApertureEfficiency

Root cause: The root cause is that the class herschel.spire.ia.dataset.PhotApertureEfficiency (among others) is new for data processed with HCSS 12 and doesn't exist in previous versions of HIPE. Previous versions cannot open tarballs containing these types of data.

Workaround: The recommended workaround is updating HIPE to version 12.1.

If you cannot upgrade HIPE, please contact the Helpdesk.

Migrating pools across incompatible versions of HIPE

HIPE 10 developer versions (builds between 2069 and 2674) and HIPE 11 (all builds) pools are incompatible with any other version of HIPE (including all public releases of HIPE 10).

In order to use these existing pools with an incompatible version, you should rebuild the index of the pool. Given that it is not guaranteed that newer data should work with older versions of the software, this guide will focus only on making the pools workable in the following cases:

  • Case 1: Using pools created/modified in a HIPE 10 version prior to 2069 with a HIPE 10 version between 2069 and 2674.
  • Case 2: Using pools created/modified in a HIPE 10 version between 2069 and 2674 with a HIPE 10 version between after 2674.
  • Case 3: Using pools created/modified in a HIPE 10 version other than the 2069-2674 range with HIPE 11.
  • Case 4: Using pools with an index rebuilt for a HIPE 10 version other than the 2069-2674 range with HIPE 11.

Prerequisite:

A pool created with a HIPE 10 version prior to 2069.

Case 1:

Pool you want to open: A pool like the one detailed in the prerequisite.

Error message: There is no error displayed and the pool is completely usable in HIPE 10 versions 2069-2674.

Case 2:

Pool you want to open: A pool that is like the prerequisite or has been processed in a scenario like CASE 1.

Error message: Querying the pool using the Product Browser returns nothing (no error message). Opening the observation from the Console using getObservation returns the following error message:

herschel.ia.task.TaskException: Error processing getObservation task: Index Version not compatible. Expected : 4 Existing: 6. 
Pool <pool_name> requires upgrading before you can use it with this software. In order to do so, you need to run pool_name.rebuildIndex() to upgrade. 
Depending on the size of the pool this process can take a long time, please be patient! 
More information can be found in the Data Analysis Guide, section 1.2.2.1 (Update of index format for local stores).

Notes about the error message:

  • pool_name is the name of the pool you are trying to load.
  • pool_name.rebuildIndex() is not the proper command to rebuild the pool. See below for the appropriate commands.
  • The section of the Data Analysis Guide referenced is not correct. When ready, a copy of this procedure will be available in the section 1.11 (Migrating pools across incompatible versions of HIPE)

Workaround: Any of the following commands will rebuild the index of the pool:

  • This ensures we are rebuilding the specified pool using a static method from the class LocalStoreFactory?:
LocalStoreFactory.getStore(pool_name).rebuildIndex()
  • This rebuilds the pool using a ProductStorage? instance named store:
store.writablePool.rebuildIndex()

Case 3:

Pool you want to open: A pool that is like the prerequisite.

Error message: There is no error displayed and the pool is completely usable in HIPE 11.

Case 4:

Pool you want to open: A pool which index was rebuilt following CASE 2 instructions.

Error message: There is no error displayed and the pool is completely usable in HIPE 11.

That's it! You should now be able to use your pool in any of the above cases. If you find you cannot, contact the Helpdesk.

tip Note that the index files will not be backed up while you run rebuildIndex(), unless you invoke Configuration.setProperty("hcss.ia.pal.pool.lstore.index.backup", "true") first. For more information execute the following command: help('dag:sec-lstore-index-compatibility'). See the HIPE Owner's Guide: Section 4.

Older HIPE versions

Known issues for older HIPE versions are listed in a separate page.

General HIPE startup issues

HIPE does not start at all, without giving error messages. What's wrong?

This can happen for different reasons, that are explored in the next entries. When you start HIPE by clicking on its icon and a problem occurs, you may only see a terminal/command window appear for a fraction of a second, leaving you without information on the error.

To obtain more information about the error, try invoking HIPE directly from a terminal/command window ([installation_dir]/apps/hipe) and examine the text output. Note that the output from some errors can be quite lengthy, so you may need to scroll up to discover the root cause.

The next entries list some common problems that prevent HIPE from starting.

HIPE cannot start: the JVM cannot reserve enough memory

This only applies if you installed HIPE using the installer, and not downloading a developer build from the Continuous Integration System.

The cause is that the Java virtual machine is unable to start with the designated amount of memory.

The terminal/command window from which you invoke HIPE shows the following error message that indicates insufficient memory:

Error occurred during initialization of VM
Could not reserve enough space for object heap
Could not create the Java virtual machine.

To fix this error you have to modify two properties in the installed.properties file within your HIPE installation directory. There is no need to reinstall HIPE.

The properties you need to modify are the following. You have to lower the amount of allocated memory:

java.vm.memory.max
java.vm.options=-XX:PermSize

Possible values are the following:

java.vm.memory.max=1024m
java.vm.options=-XX:PermSize=128m

Try running HIPE after modifying the values until it successfully starts.

ALERT! You may have the same properties defined in a hipe.props or user.props file, located in the .hcss directory within your home directory. In this case, any values set in hipe.props override those in user.props, which override those in installed.properties. Make sure to edit the properties in the file with highest proprity for your changes to take effect. See this page for more information on property files and their precedence order.

HIPE cannot start: SAXNotRecognizedException

Ticket number: HCSS-7112

The terminal/command window from which you invoke HIPE shows a SAXNotRecognizedException.

To solve this issue, add the following line to the installed.properties file in your HIPE installation directory:

java.vm.options = -Djavax.xml.parsers.SAXParserFactory=com.sun.org.apache.xerces.internal.jaxp.SAXParserFactoryImpl

HIPE cannot start: exception access violation.

Ticket number: HCSS-7139

The terminal/command window from which you invoke HIPE shows an EXCEPTION_ACCESS_VIOLATION error related to a class in org.apache.xerces.

Restart HIPE to resolve the issue.

This error is probably due to a bug in the Java Virtual Machine, and has so far been found only on Windows XP 32 bit.

I get startup messages about missing HIFI/PACS/SPIRE modules. Is that serious?

You are probably referring to messages such as the following (for HIFI; messages for PACS and SPIRE are the same):

Import "from herschel.hifi.all import *" returns:
          Import Error -> No module named hifi

Such messages are harmless and just reflect the fact that you have not installed the software related to the specified instruments.

Java 7 issues

In a Windows domain-joined computer (or with special folder redirection for user profiles) Java can be unable to write to the user home directory

Description: When executing HIPE in the context of a user that has redirected shell folders, environment variables in the Desktop path in the Registry or doesn't have a roaming profile for the domain, the Java VM cannot determine the working directory for the user or, as it is called in Windows parlance, "home path". Instead, it uses the "user profile" directory, which can be shared between several users, be read-only or even can be unresolvable for Java because it contains environment variables. In those cases, writing to this home directory (in Java terms "user.home") can fail with an error message like the one shown below.

Error message: HIPE writes a WARNING message when Java is unable to resolve the correct home directory for the user. Other error messages can be displayed and they are related to HIPE-managed pools like MyHSA, preferences, automatic backup files or anything that it is written automatically to the .hcss directory inside the home dir.

18-Mar-14 11:41:16.968 WARNING LogView : Could not log to file:
java.io.IOException: Couldn't get lock for C:\Users\wrongdir\.hcss\apps\hipe\logs\hipe_user_20140318_114116.log
       at java.util.logging.FileHandler.openFiles(Unknown Source)
       at java.util.logging.FileHandler.(Unknown Source)
       at herschel.ia.jconsole.views.LogView.(LogView.java:119)
       at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
       at sun.reflect.NativeConstructorAccessorImpl.newInstance(Unknown Source)
       at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(Unknown Source)
       at java.lang.reflect.Constructor.newInstance(Unknown Source)
       at java.lang.Class.newInstance(Unknown Source)
       at herschel.share.util.ObjectUtil.newInstance(ObjectUtil.java:153)
wrongdir is the directory that Java has retrieved using the faulty algorithm described below. user is the Windows username of the current user. In a "normal" set-up, these two strings should match.

Root cause: Due to a bug in Java, probably present since the first version, the Virtual Machine uses a non-supported method for determining the user home directory (which is retrieving a well known path from the Registry: the Desktop path, and considering the base directory as home directory). This usually works for single user, or multiple user computers which are not joined to a Windows domain, but causes errors if the computer is domain-joined, has special configuration for the redirection (symbolic links) of user directories or any other non-standard configuration related to user profiles.

Workaround: If the special folder redirection is not applied using group policies but it is a persistent change of configuration (perhaps added to the system image used to install Windows), you can change the path of the value Desktop to a user-writable directory. This value is inside the following Registry key:

HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Shell Folders

NOTE: Please take care editing the Windows Registry and be sure to follow the Microsoft recommended procedure.

Application blocked by security settings when trying to launch external Web Start applications that lack the proper publisher certificate (e.g. Aladin)

Description: As part of the security overhaul for the latest versions of Java 7, Oracle has also introduced a mechanism in which Web Start applications that are not signed with the proper publisher certificate or are self-signed are blocked. See the screenshot below in Error message.

Error message: When sending images to Aladin (which currently is self-signed), the following modal dialogue appears:

Screen_Shot_2014-01-22_at_11.09.14_AM.png

Workaround: General instructions on how to override the security mechanisms set in place by Oracle in the latest versions of the JRE are described here. Our recommendation is that you add the site from which the .jnlp file of Aladin is downloaded to the Java Exception Site List. These are the steps:

  • Open the Java Control Panel (inside Control Panel in Windows, inside System Preferences in OS X, depends on the distribution in Linux).
  • Click on the Security tab.
  • Inside the Exception Site List panel, click on the Edit Site List... button.
  • Click Add in the dialogue that opens.
  • Enter the root URI for the Aladin site (http://aladin.u-strasbg.fr).
  • An alert is displayed telling you that trusting HTTP sites is a security risk. Please click Continue to acknowledge the alert and complete the addition.

The Web Start applications downloaded from the sites you add in this way will not be subject to security checks.

Cannot find Java Runtime Environment (JRE) when launching external Web Start applications from HIPE on Mac OS X

Description: In certain combinations of Mac OS X 10.7 and Java 7, you might get an error message when trying to open the HSA user interface from HIPE. In that case, an error appears (see Error Message below) that states that there is no Java Runtime Environment installed. Even installing again, the HSA does not execute correctly.

Ticket number: HCSS-18932

Error message: The error displayed is similar to this screenshot:

Screen_Shot_2014-01-15_at_12.19.51_PM.png

Root cause: This error is due to a change in the repository directory for the Java Runtime Environment and the Java Developer Kit distributions from previous versions. References to older versions of the plug-in are kept in the system and conflict when launching Java Web Start applications. This issue is related to the one below.

Workaround: The simple workaround is to:

  1. Completely remove java 7 from the system following the instructions in http://www.java.com/en/download/help/mac_uninstall_java.xml and http://docs.oracle.com/javase/7/docs/webnotes/install/mac/mac-jdk.html
  2. Then reinstall the Java Developer Kit 7 (JDK) and Java Runtime Environment 7 from a Firefox browser following the instructions at http://www.java.com/en/download/help/mac_install.xml

Java dialogue pops up after closing an external application launched from HIPE using Java Web Start on Mac OS X

Description: When you close an external application opened from HIPE (like the Herschel Science Archive, VOSpec, Aladin, etc.), the dialogue in section "Error message" below pops up.

Error message: The following modal dialogue is displayed after closing the external application:

Screen_Shot_2014-01-27_at_10.23.53.png

Root cause: This is a known bug (although is not available at Oracle's bug database at the moment, it is in OpenJDK bug database) of Java on Mac OS X that will be fixed in Java 7 Update 60. The error is apparently caused by the presence of an alternative JavaLaunchHelper implementation in a dynamic library (.dylib extension) provided by a third party and registered as a Safari browser plug-in.

Workaround: There is no workaround at the moment (see below in "Side effects") except waiting for the new update from Oracle.

Side effect: This message does not actually indicate an error, as the external application was correctly launched, used and closed. The modal dialogue is displayed because HIPE gathers any console (error) output when the external application is executing and displays it in a modal dialogue when it is closed.

Warning appears in the console when running HIPE with a user that is not administrator in Windows 8 and later

Ticket number: HCSS-18807

Error message: This appears in a Command Prompt when starting HIPE.

WARNING WindowsPreferences <init>: Could not open/create
prefs root node Software\JavaSoft\Prefs at root 0x80000002. Windows RegCreateKey
Ex(...) returned error code 5.

Root cause: In a Windows installation with Java 7 version 21 or later, the Windows Preferences classes in Oracle's JVM try to read from the Windows Registry key: HKEY_LOCAL_MACHINE\Software\JavaSoft\Prefs . Since the key has not been created by the Java installer (as was the case in previous versions of Java), these classes try to create them and fail due to insufficient privileges. This is displayed on the Java console as seen above in the warning message.

Workaround: This warning message can be safely ignored. But if you want to remove it completely the easiest workaround is starting HIPE at least once as administrator.

Side effects: The attempt to read this Registry key does not affect the behaviour of HIPE, since HIPE stores the user preferences in the CURRENT_USER hive of the Windows Registry.

Crash when a HIPE session is left unattended for a long time with Java 7 on Mac OS X, after any display change like: setting resolution, configuring additional monitors or even switching applications after resuming work

Configuration prone to cause the error

  • Operating system: Any version of Mac OS X.
  • Java version: any Java 7 version.

Usage prone to cause the error

  • Leave HIPE running unattended for several hours.
  • Return to work with HIPE and:
    • Switch applications.
    • Wake the computer/monitor from sleep.
    • Change the display resolution.
    • Send an application to another monitor in a multimonitor setup.

Error message

The crash displays a long error trace that begins with:

Thread 0 Crashed:: AppKit Thread  Dispatch queue: com.apple.main-thread
0   com.apple.CoreFoundation         0x00007fff8d7e2a1b CFRelease + 27
1   liblwawt.dylib                   0x000000011b557647 __Java_sun_lwawt_macosx_CWrapper_00024NSObject_release_block_invoke_1 + 91
2   JavaNativeFoundation             0x000000011a8ea5f5 +[JNFRunLoop _performCopiedBlock:] + 20
3   com.apple.CoreFoundation         0x00007fff8d8a7d9d +[NSObject performSelector:withObject:] + 61
4   com.apple.Foundation             0x00007fff85d9dd70 __NSThreadPerformPerform + 214
5   com.apple.CoreFoundation         0x00007fff8d7ec4f1 __CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE0_PERFORM_FUNCTION__ + 17
6   com.apple.CoreFoundation         0x00007fff8d7ebd5d __CFRunLoopDoSources0 + 253
7   com.apple.CoreFoundation         0x00007fff8d812b49 __CFRunLoopRun + 905
8   com.apple.CoreFoundation         0x00007fff8d812486 CFRunLoopRunSpecific + 230
9   com.apple.HIToolbox              0x00007fff8acda2bf RunCurrentEventLoopInMode + 277

General information

If you are affected by this problem, it is useful to know that:

Workarounds

From the information above, the only identified workarounds for those of you that run long sessions of HIPE are:

Cursor position in the editor view is inaccurate in high-DPI (Retina) displays using Java 7 version

Ticket number: HCSS-18698
Versions affected: HIPE 12 or older
Description: The cursor position appears inside characters instead of between them.

Workaround

From Java 7 Update 55, Oracle has fixed the issue. As the bundled JRE for HIPE 12 or older is Java 7 Update 45, to be able to benefit from the patch you must change the script that launches HIPE and point it to the Java 7 Update 55 or later installed at system level. To do that, edit the file HIPE_DIR/bin/_launch.sh and edit or delete the line(s) that export the PATH variable:

PATH=/Applications/hipe_v12.0.0/jre/Contents/Home/jre/bin:$PATH
export PATH

NOTE: You should only delete the lines if you are completely sure that the default version of the system is Update 55 or later. In order to check if that is the case, you can execute java -version or /usr/libexec/java_home -V on the Terminal and then, after confirming the default version, use that path to update the _launch.sh file.

General Help System issues

When Google Chrome is the default browser for Mac OS X, the navigation arrows at the top of the help pages are not rendered

After repeated uses of the help pages for HIPE with Chrome, the images for the navigation arrow icons located in the header of the pages are (supposedly) retrieved from the browser cache but not rendered at all.

Workaround: The current workaround for this behaviour is to clear Google Chrome's cache, following these instructions.

The help does not work anymore! What's wrong?

First, check if your issue is that described below, in Why is the HIPE Help System suddenly unable to connect to any help page? If not, you have probably encountered a bug. The following are known issues that can prevent you from viewing the documentation, resulting either in a blank page or an error message:

  • A caching bug in Safari 4.0 can be solved by clearing the browser cache. Alternatively, open the help page with another web browser: copy and paste the address (something like http://127.0.0.1:8082/index.jsp?mark=null) to the address bar of the other browser.
  • A problem with Jetty's file management (Jetty is part of the help system) can result in error messages when viewing the documentation of any installed HIPE version. In particular, you may get an HTTP ERROR: 404 NOT FOUND message in your browser, and a message about index.jsp not found in the command line terminal. There are two options to solve the problem:
    • Delete the temporary directory created by the help system in /private/var/folders/[1]/[2]/T/yahajs_[username]_[HIPE version]/jetty/instance_[instance number]/webapp (the first ellipsis [1] is a pair of alphanumeric characters and the second [2] is a string of 30 alphanumeric characters). For Linux, the path will be something like /tmp/yahajs_[username]_[HIPE version]/jetty/instance_[instance number]/webapp depending on the distribution.
    • If the previous solution does not work, please try rebooting the computer.
  • Older browser versions may not display documentation correctly. For example, links from HIPE to the Help System, like the Help in URM menu entries for tasks, work in Firefox 3.6.x but not in Firefox 3.0.x. Other legacy browsers may have similar problems.
  • If you are using Linux and get an error message like Firefox can not contact the server www.%u.com, change your default browser from firefox %u to firefox %s (same if mozilla is your default browser). See the documentation of your Linux distribution for how to change default web browser. Alternatively, in HIPE choose Edit --> Preferences, go to Help & Documentation and enter the path to Firefox, or to another browser on your system, in the Browser field. Click Test to check if it works.
  • In general, when you encounter a problem with the Help System you should follow these steps:
    • See if the help works with another browser. If possible, upgrade your browser to the latest version.
    • Clear the browser cache
    • Delete the browser cookies
    • Reboot the computer
If none of the above works, please raise a ticket with the Helpdesk.

Is there a way to open the HIPE Help System without starting HIPE?

Yes, via the hipe_help command, which resides in the apps folder in your HIPE installation. The hipe_help command will start a standalone Help System in a window called Yet Another JAva Help System. Press the Display button to open the Help initial page in your default browser.

The Help System opens in my web browser. Does it mean I need to be connected to the Internet?

No, the Help System resides on your computer and does not need Internet connection, except of course for links to online resources, such as YouTube videos and pages hosted on the Herschel Wiki website, including the What's New and Known Issues documents.

Why does searching the documentation take such a long time?

When you execute the first search with a new HIPE version, HIPE has to index the contents of the documentation. This can take up to several minutes, depending on the performance of your system. Indexing of the documentation is only done for the first search. Subsequent searches will be much faster.

We are investigating ways to include pre-built indexes in the documentation shipped with HIPE, so you will not have to wait for the results of your first search.

Why is the HIPE Help System suddenly unable to connect to any help page?

The Help System relies on a local web server to display help pages. If you close the application that started the web server, the server will shut down as well:

  • If you started the Help System from HIPE, quitting HIPE will stop the Help System.
  • If you started the Help System with the show_help command, closing the Yet Another JAva Help System window will stop the Help System.

Note that when the Help System stops, any browser window or tab showing help pages will remain open, but links on them will not work anymore.

Why are several important packages missing from the Javadoc?

You are probably looking at the static Javadoc located in the doc/api/ folder of your HIPE installation. The complete Javadoc can now be accessed from the HIPE Help System by clicking on HCSS Developer's Reference Manual (API) in the Developer Reference section. The plan is to remove the static version of the Javadoc completely.

Is it possible to have the traditional, frame-based Javadoc layout?

Javadoc pages appear within the HIPE Help System frame structure. If you click on the FRAME link on any Javadoc page, you will obtain the traditional Javadoc layout. To go back to the HIPE Help System layout you will have to use the Back button of your browser (clicking on the NO FRAMES link will not work). To avoid losing the HIPE Help System layout, you may want to right-click on the FRAMES link to open the traditional Javadoc layout into a new tab of your web browser.

How can I provide feedback/requests regarding HIPE documentation?

You can leave a comment using the box at the bottom of any documentation page in the HIPE Help System. You must be connected to the Internet to view and submit comments.

Alternatively, you can click the envelope icon on the top toolbar of any help page. This opens a new message in your email client, where you can add your feedback.

If you want your issue to be officially tracked and supported, please raise a ticket with the Helpdesk of the Herschel Science Centre.

Other help issues

  • URM categories: Some entries in the three instruments User's Reference Manuals have wrong or missing categories.
  • URM examples: Some examples in the User's Reference Manuals may not work as expected. Please use the feedback icon in the HIPE Help System to let us know about broken examples.
  • JIDE vs HIPE: The documentation shipped with HIPE may still mention JIDE in some places. JIDE was the data reduction application that came before HIPE and is no longer supported. All references to JIDE can be applied to HIPE.

Herschel Science Archive issues

How to log in with new credentials after changing the password for HSA I have changed/reset the password of the account I use to access Herschel Science Archive, but HIPE still tells me I am logged in (with the old password), even after restarting HIPE. How do I log in with the new password?

Description: HIPE stores your HSA password for you if you log in and check the "Remember Me" checkbox. If you restart you can access the HSA without having to enter your info again.

However, if you reset your password using the Automated user Password Fixer, and restart HIPE, then HIPE will still report you as being logged in, even though you can't be, because you haven't told HIPE your new password yet.

Versions affected: HIPE 11 or earlier.

Ticket number: HCSS-17907

Workaround: To work around the situation, double-click on your username in the HIPE status bar at the bottom of the window to log out explicitly. (If you are not logged in, the status bar will show a label "HSA Log-in" and you don't need to do anything).

After this, you can double-click again to log in again, using the correct password, or you can wait until HIPE prompts you to log in.

I have downloaded FITS files in a tar archive from the HSA, but they are corrupted. How can I fix them?

You may have used WinZip under Windows. WinZip has an option called TAR file smart CR/LF conversion, in the Miscellaneous tab of the Configuration dialogue window (at least up to version 15.0), that is enabled by default and causes the corruption. Disable the option to solve this problem. You can also use other software such as 7-Zip.

In the HSA user interface I can query and select observations, but I do not get the option in the HSA to pass the data to HIPE, or in HIPE to load the selected products. Why?

If you are using Windows, the problem might be that you instructed your browser to auto-detect proxy settings. In Windows this is a system-wide setting which affects any application connecting to the Internet, including HIPE.

You can change proxy settings for HIPE by following the instructions in the HIPE Owner's Guide.

When I click Send to external Application in the HSA interface, nothing seems to happen. What is going on?

The data is being transferred in the background. Clicking Send to external Application should bring to the front the main HIPE window, which will tell you that data is indeed being received. In some cases however the HIPE window does not appear. Bring it to the front manually to follow the progress of your action.

When I try to retrieve data from the HSA I get an error message in the console about an invalid magic number for FITS file. What is going on?

The reason is that the "data" being downloaded is in reality a small file containing an error message. Two issues have been shown to cause this problem:

  • Your HSA password has non-alphanumeric characters in it. Change it to a password containing only letters and numbers.
  • Your HSA username has uppercase characters. Write it using only lowercase letters.

Other issues

HIPE crashed/froze! What do I do?

If you cannot find a solution in this document, you may have found a bug in HIPE. Please report it by opening a Helpdesk ticket. If the system produced a dump file when the crash happened, make sure to include it in your ticket. For more information about dump files and where to look for them, see the HIPE Owner's Guide.

Cannot open a FITS file created by HIPE

If you export a FITS file from HIPE and modify it with an external program, HIPE may not be able to import it anymore. If this happens, follow these steps:

  1. Open the FITS file with a FITS editing program such as fv.
  2. Delete the HCSS____ keyword from the header of all extensions.
  3. Save the file.

HIPE should now be able to read the file.

Cannot run WebStart applications on Mac OS X Lion or Mountain Lion because of security settings

WebStart applications that can be opened from within HIPE include the HSA User Interface, Aladin, VOSpec and so on. When trying to open one of such applications, a window like the following may appear:

topcat.png

This is caused by the security settings of the Gatekeeper feature of Mac OS X. For more information about Gatekeeper, see the Apple support website.

To allow a WebStart application that is blocked, change the Allow applications downloaded from option in Gatekeeper from Mac App Store to one of the two less restrictive options. If the application has a valid certificate, the Mac App Store and identified developers option will suffice. Otherwise you will need to select the Anywhere option.

warning The Gatekeeper feature only handled Java applications for the Apple-specific version in older Mac OS X (before Mac OS 10.7.2). The Java version from Oracle has a mechanism for validating publisher certificates completely independent from the operating system (see the related archived known issue for the HSA UI).

"Invalid context" errors on Mac OS X Lion

When you start HIPE form a Terminal window on Mac OS X Lion, you may see errors such as these:

java[40745] <Error>: CGContextGetCTM: invalid context

These errors are harmless and can be safely ignored.

Problems saving observations to NFS-mounted disks and FAT32 disks

When saving observation to your computer, you may experience problems in the following cases:

  • When saving data to NFS-mounted disks (you may get an IllegalMonitorStateException).
  • When saving data to a FAT32 filesystem from a Mac (you may get an OverlappingFileLockException).
Both problems are solved by setting the following property:
hcss.ia.pal.pool.lstore.lock = simple
See the HIPE Owner's Guide for information on how to set properties.

Why do I get a ClassCastException when trying to save an observation with the saveObservation command?

You probably retrieved the observation without having the corresponding instrument software installed (for instance, you retrieved a PACS observation without having the PACS modules in HIPE).

To solve this problem follow these steps:

  1. Reinstall HIPE, including the software of the desired instrument.
  2. Delete the following directories (home being your home directory):
    • home/.hcss/lstore/hsa_cache
    • home/.hcss/pal_cache/hsa
  3. Retrieve the observation again.

I get an OutOfMemoryError related to PermGen space. What can I do?

Add or edit the following property in your hipe.props property file:

java.vm.options = -XX:MaxPermSize=256m

Note that this parameter is specific to the Sun/Oracle Java implementation, which is the only one officially supported by HIPE.

For more information about property files, see the HIPE Owner's Guide.

There are PACS files I cannot delete! Why?

It might not be possible to delete the following files on Windows because their full path exceeds the 260 characters limit. This is caused by a limitation in Windows Explorer and the Command Prompt to maintain backwards compatibility (even if the operating system supports longer file paths).

  • PCalSpectrometer_RelCalSourceFlux_FM_v1.fits_2009...80Z.fits
  • PCalSpectrometer_RelCalSourceFlux_FM_v2.fits_2009...55Z.fits
  • PCalSpectrometer_RelCalSourceFlux_FM_v2.fits_2009...14Z.fits
  • PCalSpectrometer_RelCalSourceFlux_FM_v2.fits_2009...21Z.fits
  • PCalSpectrometer_RelCalSourceFlux_FM_v2.fits_2009...58Z.fits

You can avoid this problem by installing HIPE with a short initial path, such as C:\hipe, so that full paths remain below the limit.

If you already have several files with long paths that you wish to delete, you can delete them from the Command Prompt (CMD.EXE) prepending the path with \\?\. For example for deleting C:\too\long\path\fitsfilename.fits you should write:

C:\>del \\?\C:\too\long\path\fitsfilename.fits
and press Enter.

HIPE does not accept keyboard input anymore! What's wrong?

If you are running Linux, this could be due to a compatibility problem between Java and the input method in use, like IBus or SCIM (Smart Common Input Method). A workaround is to stop the service of the input method or uninstall it. Please refer to the documentation of your distribution for how to do so.

The problem is described in more detail on these pages:

What are all the tmp_product_sink_* pools in my local store directory?

You may find these pools in your local store directory (by default .hcss/lstore in your home directory) after reducing data with HIPE. These are temporary pools that are not deleted upon quitting HIPE due to a known bug. You can safely delete them by hand after quitting HIPE.

toggleopenShow attachmentstogglecloseHide attachments
Topic attachments
I Attachment Action Size Date Who Comment
pngpng HIPE_addPools.png manage 39.2 K 19 Jul 2010 - 15:08 PaulBalm Screenshot for: PAL Storage Manager: Can't add pools to storages
pngpng java7u45-font-issue.png manage 11.3 K 21 Nov 2013 - 09:07 AlvarGarcia Java 7 update 45 font issue
pngpng Screen_Shot_2014-01-16_at_16.28.23.png manage 48.3 K 17 Jan 2014 - 10:00 AlvarGarcia Java 7 Permissions security feature
pngpng Screen_Shot_2014-01-16_at_16.29.26.png manage 89.3 K 17 Jan 2014 - 10:03 AlvarGarcia Java 7 Permissions error dialog
pngpng Screen_Shot_2014-01-27_at_10.23.53.png manage 33.1 K 27 Jan 2014 - 13:07 AlvarGarcia Java 7 WebStart? message
pngpng Screen_Shot_2014-01-15_at_12.19.51_PM.png manage 282.3 K 30 Jan 2014 - 10:47 AlvarGarcia Java 7 corrupted installation
pngpng Screen_Shot_2014-01-22_at_11.09.14_AM.png manage 29.0 K 14 Feb 2014 - 08:46 AlvarGarcia Java 7 self-signed WS applications
This topic: Public > WebHome > HipeKnownIssues
History: r165 - 21 Oct 2014 - 15:18:58 - AlvarGarcia