http://www-glast.stanford.edu/protected/mail/opsprob/
Area |
Problem |
Comments |
Resolution |
---|---|---|---|
Monitoring |
Comparing Histograms |
It is not currently possible to compare the same histogram from different runs |
|
Login |
Login in GLAST data portal |
One user was not registered in the glast user database, several users had missing SLAC username's |
Fixed |
Tomcat |
alarms summary doesn't work |
Connection broken between IIS and server. Not a crash |
The tomcat server was restarted. A page has been created for tracking similar outages http://confluence.slac.stanford.edu/x/qQFV |
Data Monitoring |
Alarm summary still not working |
Same as above |
|
ASP |
overlapping bars in ASP flares |
A data representation issue. |
Fixed |
Glast Ground |
Menu item for shifts had not been updated since Ops Sim 1 |
|
Fixed |
Data Monitoring |
Problems with Alarm documentation |
No readout errors in MC |
|
Data Access |
skimming on single runids gives bad files |
It is currently not easy to skim OpsSim2 data because of multiple versions of the same files. |
|
Data Access |
Problems using download manager with CAL files |
Cal files are getting negative file sizes. Download managed chokes on these. |
Several JIRA's created: GDCB-49@JIRA DATASERV-104@JIRA |
Data Monitoring |
plot not showing the value that triggered the alert? |
Trending plots are accumulated by default. Usability issue. |
|
Data Monitoring |
slow making plots |
Big data request, many users, db overloaded |
|
Data Monitoring |
Data processing files inaccessible |
Connection broken between IIS and server. Not a crash. |
The server was restarted and the error recorded at http://confluence.slac.stanford.edu/x/qQFV |
Data Monitoring |
Data Quality Monitoring Page stalled |
Same as above. |
|
Data Monitoring |
missing plots |
Not sure which plots |
|
Data Monitoring |
Monitoring seems to be down again... |
Due to missing data |
|
ASP |
Problem with GCN notices |
|
|
Documentation |
Who is on shift in Confluence |
Link in confluence page was broken |
Fixed |
ASP Data Viewer |
ASP Start time and End time |
Bug |
Fixed |
ASP Data Viewer |
Can not access page directly |
User had to login. Usability issue. |
|
Data Monitoring |
TowerCalLayerCalColumnCalXFace plots not being displayed |
Some users cannot see plots. |
Under investigation. |
ASP |
Source Monitoring |
User was using wrong tool |
User instructed to use http://glast-ground.slac.stanford.edu/ASPDataViewer SM-20@JIRA |
Pipeline |
GLAST PIpeline II PAge |
Request to make it easier to identfy failed runs |
JIRA created: PFE-156@JIRA |
Data Monitoring |
Missing plots |
Recon_ReconAcdPhaMips_PMTA_Zoom_TH1_AcdTile metadata type was incorrect (seems to have changed during the day?) |
Fixed |
Data Monitoring |
Ingestion problem |
Some data not getting ingested. Some data seems duplicated. |
Under investigation. |
Area |
Problem |
Comments |
Resolution |
---|---|---|---|
Data Monitoring |
Processing errors |
Caused by problems in trending ingestion code |
Fixed? |
JIRA |
Exception on JIRA homepage |
"Bad file descriptor" in Lucene search engine |
JIRA (and confluence) server was restarted |
Data Monitoring |
Trending Configuration "Max Number of bins" |
Setting Max Number of bins=10 has no effect |
|
L1Proc |
Fast mon crash for run 258280320 |
|
|
L1Proc |
Processing error for run 258292096 |
bug in gcrCalib that occasionally makes recon segfault. |
Fixed by rollback |
Logging |
Logs for 80306007 |
Log reports were missing from data processing page after processing complete |
The application that posts messages to the database from the central message file was having a problem and had to be restarted |
Cross Trending |
Scatter plots |
Scatter plots for the cross trending are not always displayed |
Under investigation |
Data Access |
Problem accessing root files from noric |
|
Fixed by using root 5.18 ? |
Pipeline |
pipeline mail processing |
Mail delivery to pipeline server sometimes slow |
Under investigation |
Data Monitoring |
Incorrect/Slow plot display |
|
A bug in selection of histograms for vector quantities has been found and fixed (but not deployed). Some progress has been made on understanding poor performance of plot displays, but problem remains under investigation |