properly working stale handle detection

release hltd 1.6.2-2 and fffmeta 1.6.2-0
cmsdaq · Mar 5, 2015 · e19ef6b · e19ef6b
1 parent d381173
commit e19ef6b
Show file tree

Hide file tree

Showing 4 changed files with 7 additions and 7 deletions.
diff --git a/python/elasticbu.py b/python/elasticbu.py
@@ -257,7 +257,7 @@ def elasticize_box(self,infile):
             #only here
             document['host']=basename
             try:
-                document['detectedStaleHandle']=bool(document['detectedStaleHandle'])
+                document['detectedStaleHandle']=bool(document['detectedStaleHandle']=='True')
             except:
                 pass
             self.index_documents('boxinfo',[document])

diff --git a/python/hltd.py b/python/hltd.py
@@ -545,7 +545,7 @@ def run(self):
                                 mpstat = os.stat(disk)
                             #no issue if we reached this point
                             fu_stale_counter = 0
-                        except IOError as ex:
+                        except (IOError,OSError) as ex:
                             #TODO:which kind of error is thrown with unresponsive Force10 network
                             if ex.errno == 116:
                                 if fu_stale_counter==0 or fu_stale_counter%20==0:
@@ -587,10 +587,10 @@ def run(self):
                                 numQueuedLumis,maxCMSSWLumi=self.getLumiQueueStat()
                                 fp.write('activeRunNumQueuedLS='+numQueuedLumis+'\n')
                                 fp.write('activeRunCMSSWMaxLS='+maxCMSSWLumi+'\n')
-                                fp.write('detectedStaleHandle='+str(fu_stale_counter>0))
+                                fp.write('detectedStaleHandle='+str(fu_stale_counter>0)+'\n')
                                 fp.write('entriesComplete=True')
                             boxinfo_update_attempts=0
-                        except IOError as ex:
+                        except (IOError,OSError) as ex:
                             logger.warning('boxinfo file write failed :'+str(ex))
                             #detecting stale file handle on recreated loop fs and remount
                             if conf.instance!='main' and (ex.errno==116 or ex.errno==2) and boxinfo_update_attempts>=5:
@@ -602,7 +602,7 @@ def run(self):
                                 time.sleep(1)
                             boxinfo_update_attempts+=1
                         except Exception as ex:
-                            logger.warning('boxinfo file write failed +'+str(ex))
+                            logger.warning('exception on boxinfo file write failed : +'+str(ex))
 
                     if conf.role == 'bu':
                         #ramdisk = os.statvfs(conf.watch_directory)

diff --git a/scripts/hltdrpm.sh b/scripts/hltdrpm.sh
@@ -196,7 +196,7 @@ cd $TOPDIR
 cat > hltd.spec <<EOF
 Name: hltd
 Version: 1.6.2
-Release: 1
+Release: 2
 Summary: hlt daemon
 License: gpl
 Group: DAQ

diff --git a/scripts/metarpm.sh b/scripts/metarpm.sh
@@ -168,7 +168,7 @@ cd $TOPDIR
 # we are done here, write the specs and make the fu***** rpm
 cat > fffmeta.spec <<EOF
 Name: $PACKAGENAME
-Version: 1.6.1
+Version: 1.6.2
 Release: 0
 Summary: hlt daemon
 License: gpl