 |
 |
The following job (see below) received an ECANCELED error. In this particular case, the server was tripped up by sendfile() returning 0 bytes sent which the server interpreted as being recoverable but failed to recover and passed that error back to the client. The SEGV was due to insufficient error checking in the user's application. That said, the server should properly recover from this, albiet rare, situation.
Relevant snippet from the client's log:
I am using Doug Benjamin's SMWZ example to access files using a
STORAGEPREFIX of root://fax.mwt2.org:1094/. I am seeing this failure which
ultimately leads to a segmentation fault:
EvtBuilder::Notify: Setting chain offset to 311087
Info in EvtReader::Notify: Number of entries per file - 51550.
EvtBuilder::Notify: Setting file_counter to 9
140506 06:26:23 14393 Xrd: CheckErrorStatus: Server [xrdfed08.cern.ch:1094]
declared: Unable to read
/atlas/rucio/data12_8TeV:NTUP_SMWZ.01122387._000099.root.1; operation
canceled(error code: 3005)
Error in TBasket::Streamer: The value of fObjlen is incorrect (-853729090) ;
trying to recover by setting it to zero
.....Many such lines...
SEGV
—
Reply to this email directly or view it on GitHub.
Use REPLY-ALL to reply to list
To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1