Bugs item #2407834, was opened at 2008-12-09 00:09
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2407834&group_id=56967

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: PF/runtime
Group: Pathfinder "stable"
Status: Open
Resolution: None
Priority: 7
Private: No
Submitted By: Peter Boncz (boncz)
Assigned to: Nobody/Anonymous (nobody)
Summary: logger-rcopy BBP bug

Initial Comment:
pathfinder.mx:
1671     ws.slice(NID_RID,ATTR_PROP).[append](mem);

I saw a just created rcopy()-ed PROP_VAL bat from an updatable multi-doc 
collection cause an assert() in CMD_append, as it got unloaded from the BBP (ie 
BBP[].b[0] == NULL). I observed it having both a logical *and* a physical BBP 
refcount of 1. Question is how it could be unloaded. A strange clue was its BBP 
status (4, that is BBPTMP)

BBPTMP are bats that are not persistent (that is correct, it is a rcopy) but 
that got saved. That is strange, because why would that bat get saved? Note 
that an rcopy() bat is a "logical view" (lview) on the master bat. 

Another clue was that the logger background thread was in the middle of a 
BBPsync() when this assert went off. Regrettably I lost the gdb session so I 
could not debug further.

possble ideas:
- maybe the new logger somehow gets to checkpoint the copied private images 
instead of the master bats?
- meybe the rcopy bats are not properly recognized by the BBPtrim thread such 
that it saves and unloads them.



being unloaded in BBP,  

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2407834&group_id=56967

------------------------------------------------------------------------------
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs

Reply via email to