Opened 5 months ago

Closed 4 months ago

#1542 closed defect (fixed)

rasserver produces too many log files sometimes

Reported by: dmisev Owned by: dmisev
Priority: critical Milestone: 9.4
Component: rasmgr Version: development
Keywords: Cc: pbaumann
Complexity: Medium


In certain cases it seems that rasdaman produces an insane amount of log files. So many that

rm -f *
-bash: /usr/bin/rm: Argument list too long

I think rasservers die before they properly start, and rasmgr keeps forking new ones for way too long.

We need to figure out how to reproduce this first, and probably rasmgr should be fixed, so that it detects such an issue a bit earlier than ~400,000 restarts later. To reproduce we could try inserting a deliberate segfault somewhere early in the initialization of rasserver.

Change History (3)

comment:1 Changed 5 months ago by dmisev

  • Owner changed from bbell to dmisev
  • Status changed from new to assigned

comment:2 Changed 5 months ago by dmisev

To fix this I modified to try a simple select query after starting rasservers; if it fails to execute it means that no rasserver has probably started (a select version() query should always execute fine otherwise). In this case it shuts down rasmgr and hence prevents it going into the above loop of constantly restarting the rasservers.

Patch is on the way.

comment:3 Changed 4 months ago by dmisev

  • Resolution set to fixed
  • Status changed from assigned to closed
Note: See TracTickets for help on using tickets.