Bugzilla – Full Text Bug Listing
| Summary: | (Plat) No data collection when using network drives. | ||
|---|---|---|---|
| Product: | z_Archived | Reporter: | Heiner Westphal <westphal> |
| Component: | Hyades | Assignee: | Nellie Chau <nelliec> |
| Status: | CLOSED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | Brian_Young |
| Version: | unspecified | Keywords: | Documentation |
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Windows XP | ||
| Whiteboard: | closed460 | ||
BTW. I use Windows XP Home SP2 eclipse 3.0.1 hyades.rac.nt_3.2.0_20041101_1032 and the other hyades*_3.2.0_20041101_1032 without .rac. The rac is running as service. JVM is Sun J2SDK 1.4.2_04. We will investoigate this in the next iteration. I know we have looked into this in the past without any luck. It appears that the environment settings to logical network drives are not honoured by windows. There are 2 scenarios which might show different behaviour. Case 1: RAServer is running as a Windows system service The RAServer is configured to run as system account and may not have the required rights to access the network resource. In order to allow the RAServer service to access the network drive, you will need to edit the service properties. Under the "Log On" page, change from "Local System account" to "This account" and fill out the user name and password which is able to access the network drive. Restart RAServer service and you should be able access the network drive now. Case 2: RAServer is running as a standalone process If RAServer is started from a command prompt, it will inherit the current user setting and should have no problem accessing the network drive if the current user is able to do so. I have encountered this same problem in RAD v6 (based upon Eclipse 3.0). I am in a similar situation as Case 1 in comment #3 above. The Agent Controller has permissions to access the network volume but I still get a no class found error. Brian, can you show me how you can check the access permission of the Local System Account? When "User A" has a network drive mounted as "T:", only him/her can see this drive letter - i.e. other users/services on this machine cannot see this drive letter. I would recommend to use the UNC name of the network resource if possible, e.g. "\\machine\share" instead of a mounted drive letter. You can verify this by telnet'ing into the same Windows machine with a differ user ID. You will not be seeing the mounted network drives by another user. I would like to close this off since this is a operating system limitation. If there is no objection then I will go ahead. Thanks. In our network environment certain network volumes are always assigned the same drive letter AND certain directories are public (read/write) to anyone who has the volume mounted. So in our case it does not matter if UserA or UserB logs into a machine, these drive letter mappings and directory access will be available to both. In this situation environment, would you say that the RAC should be able to access the network drive? Since the RAC service is not logged in as any type of user (it is a system service) it may or may not have the environment required to access the public mounted network drives. Have you check if you can use "\\machine\share" instead? I think it is up to Windows to determine whether system services can access this kind of resource or not. If the RAC is running standalone (in the user's own environment and thus can be thought of being logged in as that user) then there should be no problem using the network drives. transfer to tptp I did some more testing and data collection works fine, if RAServer is running as standalone application. The system service does not see mounted network drives AND does not know about virtual drives created via subst. Thus this is no bug in Hyades, or TPTP, but an FAQ: Q: Why does profiling fail when using network drives? A: Because your RAServer is running as service with system user privileges without access to the network drive. To use network or virtual drives RAServer MUST either a) run as application or b) as service logged in as a user who has access to the respective drives. May be it would help to improve the error message so one can see there is an access problem. Possibly the "Caused by" chain of the exception has all info needed. undo move to tptp... Ok, we can add your FAQ to the docs. FYI. The next major release of Agent Controller should have a completely new error handling mechanism. We cannot provide such an improvement in the current release. This will be added as a limitation in 3.3 docs: Data collection fails on network or virtual drives If Agent Controller is configured to run as a system service, it does not see mounted network drives or virtual drives. Therefore, data collection will fail when collecting data on network or virtual drives. To work around this limitation, the Agent Controller should be configured in one of two ways: 1. To run as application. 2. To run as service logged in as a user who has access to the respective drives. Fixed in 3.3 setting target based on last change date, please adjust if not correct As of TPTP 4.6.0, TPTP is in maintenance mode and focusing on improving quality by resolving relevant enhancements/defects and increasing test coverage through test creation, automation, Build Verification Tests (BVTs), and expanded run-time execution. As part of the TPTP Bugzilla housecleaning process (see http://wiki.eclipse.org/Bugzilla_Housecleaning_Processes), this enhancement/defect is verified/closed by the Project Lead since this enhancement/defect has been resolved and unverified for more than 1 year and considered to be fixed. If this enhancement/defect is still unresolved and reproducible in the latest TPTP release (http://www.eclipse.org/tptp/home/downloads/), please re-open. |
When the working dir resides on a network drive (fedora core 1 with samba 3.x), I cannot collect profiling data. I get an error dialog telling: IWAT0008E Error launching X. Reason: IWAT0008E Either the agent controller (RAServer.exe) is not started, or the Java path is incorrect. See log for more details. The RAServer's sericelog.log says: <SERVER_MSG time="2004:11:4:13:6:11" severity="INFORMATION" text="Attempting to launch process : java.exe -XrunpiAgent:server=controlled X in location N:\wor k\eclipse\workspace\playground" file="..\RAServer\launcher.c" line="315"/> <SERVER_MSG time="2004:11:4:13:6:11" severity="SEVERE" text="Launching process failed, platform specific error is 267. Path=D:\Programme\j2sdk1.4.2_04\jre\bin \java.exe Process command line = "D:\Programme\j2sdk1.4.2_04\jre\bin\java.exe" -XrunpiAgent:server=controlled X" file="..\RAServer\launcher.c" line="328"/> When I change the working directory to something local, e.g. C:\work\eclipse\workspace\playground, I get java.lang.NoClassDefFoundError: X. This is another bug concerning classpath external to the workspace (independent of local or network drives). Running the application without profiling works. The class X.java ist: --------------------------------------------- public class X { public static void main(String[] args) { System.out.println("Hello world"); } } ---------------------------------------------