(r13) WmsBugs3dot1dot100 < EgeeJra1It

glite_wms_R_3_1_100
- WMPROXY
- WM
- JC/LM
- ICE
- LB
- Configuration
- Other

glite_wms_R_3_1_100

WMPROXY

Command glite-wms-job-info --jdl doesn't work with nodes of a collection

[ale@cream-15 UI]$ glite-wms-job-info --jdl https://devel17.cnaf.infn.it:9000/RAeQ1KgqrjHK4MYzB8t8Hw

Connecting to the service https://devel19.cnaf.infn.it:7443/glite_wms_wmproxy_server


Error - WMProxy Server Error
Unable to read file: /var/glite/SandboxDir/RA/https_3a_2f_2fdevel17.cnaf.infn.it_3a9000_2fRAeQ1KgqrjHK4MYzB8t8Hw/JDLToStart
(please contact server administrator)

Method: getJDL

Sometimes wmproxy returns this error message submitting a job:

Connection failed: HTTP Error
<html><head>
<title>500 Internal Server Error</title>
</head><body>
<h1>Internal Server Error</h1>
<p>The server encountered an internal error or
misconfiguration and was unable to complete
your request.</p>
<p>Please contact the server administrator,
 [no address given] and inform them of the time the error occurred,
and anything you might have done that may have
caused the error.</p>
<p>More information about this error may be available
in the server error log.</p>
</body></html>

Error code: SOAP-ENV:Server

I submit a job with a proxy of a VO which is not enabled on the WMS; the error message is not so clear:

Warning - Unable to delegate the credential to the endpoint: https://devel19.cnaf.infn.it:7443/glite_wms_wmproxy_server
Unknown Soap fault
Method: Soap Error

BUG #23443: NOT FIXED
- Required documents are not put into the glite doc template in edms
- For the R6 (JDL howto) document a broken link is given

Submitting a "parametric" jobs it always failed with this error:

*************************************************************
BOOKKEEPING INFORMATION:

Status info for the Job : https://devel17.cnaf.infn.it:9000/wzOz1QuyFAEJ5JnwXHUhng
Current Status:     Aborted
Status Reason:      requirements: unable to complete the operation: the attribute has not been initialised yet
Submitted:          Wed Jun 25 16:25:03 2008 CEST
*************************************************************

To help debugging:

27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": ------------------------------- Fault Description --------------------------------
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": Method: jobStart
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": Code: 1502
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": Description: requirements: unable to complete the operation: the attribute has not been initialised yet
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": Stack:
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": AdEmptyException: requirements: unable to complete the operation: the attribute has not been initialised yet
        at Ad::delAttribute(const string& attr_name)[Ad.cpp:180]
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart":   at JobAd::delAttribute(const string& attr_name)[JobAd.cpp:292]
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart":   at submit()[wmpcoreoperations.cpp:1936]
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart":   at jobStart()[wmpcoreoperations.cpp:1105]
27 Jun, 13:42:53 -D- PID: 5326 - "wmpgsoapoperations::ns1__jobStart": ----------------------------------------------------------------------------------

The status of a collection (or a dag) is corretly set to CLEARED after a "glite-wms-get-output", but its nodes remain in "Done (Success)" status.

*************************************************************
BOOKKEEPING INFORMATION:

Status info for the Job : https://devel17.cnaf.infn.it:9000/xe-J0mjeFjU86p2mstn8-A
Current Status:     Cleared
Status Reason:      user retrieved output sandbox
Destination:        dagman
Submitted:          Wed Jun 25 16:08:43 2008 CEST
*************************************************************

- Nodes information for:
    Status info for the Job : https://devel17.cnaf.infn.it:9000/6NeMTtefAmTCRI0N1waKzA
    Current Status:     Done (Success)
    Logged Reason(s):
        - Job terminated successfully
    Exit code:          0
    Status Reason:      Job terminated successfully
    Destination:        ce-01.roma3.infn.it:2119/jobmanager-lcgpbs-cert
    Submitted:          Wed Jun 25 16:08:43 2008 CEST
*************************************************************

    Status info for the Job : https://devel17.cnaf.infn.it:9000/OpPpNrh1EYNwPhjQQaVazQ
    Current Status:     Done (Success)
    Logged Reason(s):
        - Job terminated successfully
    Exit code:          0
    Status Reason:      Job terminated successfully
    Destination:        atlasce01.na.infn.it:2119/jobmanager-lcgpbs-cert
    Submitted:          Wed Jun 25 16:08:43 2008 CEST
*************************************************************

One collection over 60 submissions remains in "Waiting" status due to a LB problem (its nodes are in status "Submitted"); this are the wmproxy logs:

 
23 Jun, 18:58:18 -D- PID: 24834 - "WMPEventlogger::isAborted": Quering LB Proxy...
23 Jun, 18:58:18 -D- PID: 24834 - "WMPEventlogger::isAborted": LBProxy is enabled
Unable to query LB and LBProxy
edg_wll_QueryEvents[Proxy]
Exit code: 4
LB[Proxy] Error: Interrupted system call
(edg_wll_plain_read())
23 Jun, 18:58:18 -D- PID: 24834 - "wmpcoreoperations::jobStart": Logging LOG_ENQUEUE_FAIL
23 Jun, 18:58:18 -D- PID: 24834 - "WMPEventlogger::logEvent": Logging to LB Proxy...
23 Jun, 18:58:18 -D- PID: 24834 - "WMPEventlogger::logEvent": Logging Enqueue FAIL event...
23 Jun, 18:58:19 -D- PID: 24834 - "WMPEventlogger::testAndLog": LB call succeeded
23 Jun, 18:58:19 -D- PID: 24834 - "wmpcoreoperations::jobStart": Removing lock...
23 Jun, 18:58:19 -I- PID: 24834 - "wmpgsoapoperations::ns1__jobStart": jobStart operation completed

23 Jun, 18:58:19 -I- PID: 24834 - "wmproxy::main": -------- Exiting Server Instance -------
23 Jun, 18:58:19 -I- PID: 24834 - "wmproxy::main": Signal code received: 15
23 Jun, 18:58:19 -I- PID: 24834 - "wmproxy::main": ----------------------------------------

There is this warning in wmproxy submitting collections:

19 Jun, 14:58:32 -W- PID: 2746 - "wmpcoreoperations::submit": Unable to find SDJRequirements in configuration file

Fond a problem in the wmproxy. After some collections submissions it stops working and with the top command you can see that the glite-wms-proxy processes are using all the CPU (It seems that the problem is related with the porting to gsoap 2.7.10):

[Thu Jun 19 15:43:50 2008] [warn] FastCGI: (dynamic) server "/opt/glite/bin/glite_wms_wmproxy_server" (pid 7938) terminated due to uncaught signal '11' (Segmentation fault)

WM

If you remove a pending node the request is not removed from the "old" directory of the jobdir:

/var/glite/workload_manager/jobdir/old:
total 16
-rw-r--r--  1 glite glite 250 Jul  9 13:37 20080709T113710.627592_3086920576
-rw-r--r--  1 glite glite 250 Jul  9 13:37 20080709T113711.225739_3086920576
-rw-r--r--  1 glite glite 250 Jul  9 13:37 20080709T113711.794866_3086920576
-rw-r--r--  1 glite glite 250 Jul  9 13:37 20080709T113712.361657_3086920576

After the recovery of a collection the "pending" nodes are forgotten.

BUG #27215: FIXED (see also #38359 FIXED)
- If the sum of the dimensions of the first two files is exactly equal to the limit, the job stays in state RUNNING for ever and ever.

BUG #38509: The WM's recovery procedure hangs if no relevant events are found for a given request

BUG #38366: Recovery doesn't work with a list-match request. FIXED

25 Jun, 15:31:50 -I: [Info] operator()(dispatcher.cpp:754): Dispatcher: starting
25 Jun, 15:31:50 -I: [Info] operator()(dispatcher.cpp:757): Dispatcher: doing recovery
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:389): RequestHandler: starting
25 Jun, 15:31:50 -I: [Info] operator()(recovery.cpp:613): recovering https://localhost:6000/4xvi9qZqMbrNgQeuzOIGbA
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:389): RequestHandler: starting
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:389): RequestHandler: starting
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:389): RequestHandler: starting
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:389): RequestHandler: starting
25 Jun, 15:31:50 -I: [Info] main(main.cpp:292): running...
25 Jun, 15:31:50 -W: [Warning] single_request_recovery(recovery.cpp:317): cannot create LB context (2)
25 Jun, 15:31:50 -E: [Error] operator()(dispatcher.cpp:779): Dispatcher: cannot create LB context (2). Exiting...
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:467): RequestHandler: End of input. Exiting...
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:467): RequestHandler: End of input. Exiting...
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:467): RequestHandler: End of input. Exiting...
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:467): RequestHandler: End of input. Exiting...
25 Jun, 15:31:50 -I: [Info] operator()(RequestHandler.cpp:467): RequestHandler: End of input. Exiting...
25 Jun, 15:31:50 -I: [Info] main(main.cpp:295): TaskQueue destroyed

JC/LM

ICE

LB

The status of the collection is not correctly computed
glite-lb-interlogd doesn't stop (you need a "kill -9")

Configuration

Add glite-wms-ice at the metapackage dependencies
Remove "rgma" fron the enviroment variable GLITE_SD_PLUGIN
Add "globus gridftp" at the gliteservices list (gLiteservices)
Remove from the metapackage these rpms:
- checkpointing
- partitioner
- interactive (and thirdparty-bypass)
- rgma-* (and service-discovery-rgma-c)
glite-lb-interlogd doesn't stop (you need a "kill -9")
Modify the cron purger file

# Execute the 'purger' command at every day except on Sunday with a frequency of six hours
3 */6 * * mon-sat glite . /opt/glite/etc/profile.d/grid-env.sh ; /opt/glite/sbin/glite-wms-purgeStorage.sh -l /var/log/glite/glite-wms-purgeStorage.log -p /var/glite/SandboxDir -t 604800 > /dev/null 2>&1

# Execute the 'purger' command on each Sunday (sun) forcing removal of dag nodes,
# orphan dag nodes without performing any status checking (threshold of 2 weeks).
0 1 * * sun glite . /opt/glite/etc/profile.d/grid-env.sh ; /opt/glite/sbin/glite-wms-purgeStorage.sh -l /var/log/glite/glite-wms-purgeStorage.log -p /var/glite/SandboxDir -o -s -n -t 1296000 > /dev/null 2>&1

Set the ICE configuration section
Remove IsmBlackList parameter
Set jobdir as the default for the DispatcherType parameter of wm
Add the parameter EnableRecovery = true; ti the wm configuration section
Add this cron to create the host proxy:

0 */6 * * * glite . /opt/glite/etc/profile.d/grid-env.sh ; /opt/glite/sbin/glite-wms-create-proxy.sh /var/glite/wms.proxy /var/log/glite/create_proxy.log

Rotate the logs of the purgeStorage
Check bugs #35244
BUG #30900: default value on WMS conf file is again: MinPerusalTimeInterval = 5;

Other

On script glite-wms-purgeStorage.sh the command glite-wms-create-proxy.sh has to be called with its path:

[glite@devel19 ~]$ /opt/glite/sbin/glite-wms-purgeStorage.sh -l /var/log/glite/glite-wms-purgeStorage.log -p /var/glite/SandboxDir -t 302400
Certificate will expire
/opt/glite/sbin/glite-wms-purgeStorage.sh: line 27: glite-wms-create-proxy.sh: command not found

Probably on /opt/glite/sbin/glite-wms-create-proxy.sh it is better to use the --force option for command mv, instead the move commands should be blocked here:

[glite@devel19 ~]$ mv /var/glite/wms.proxy.14913 /var/glite/wms.proxy
mv: overwrite `/var/glite/wms.proxy', overriding mode 0400?

-- AlessioGianelle - 20 Jun 2008

Topic revision: r13 - 2008-07-09 - AlessioGianelle

Account
- Log In

~~Edit~~
~~Attach~~