You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I filled up the hard disk; lots of things started failing, as expected.
Upon freeing up some space, the database still refused to start: "!FATAL: Logger_new: inconsistent database, catalog does not exist"
I have attached:
The contents of merovingian.log around the time of failure
The contents of BBP.dir as it was when I got to it
The output of ls -lR
By comparing (2) and (3) you can see that the sql_catalog_bid should be at 01/103.{head|tail}, but that those files do not exist. This is the immediate cause of the FATAL error above.
The SIGBUS which took down mserver5 for the first time is, I believe, fixed in more recent releases (I'm trying to get the machine upgraded, honest!).
If you believe that problems which may have caused irrecoverable corruption of this type have been fixed in more recent releases as well then feel free to close this bug with extreme prejudice. I filed this bug because I don't remember seeing any changes of this kind in release notes or checkin comments.
I don't have an core dumps for any of this stuff - they were trying to be written to the disk that was full.
Date: 2016-04-22 15:31:04 +0200
From: Richard Hughes <<richard.monetdb>>
To: GDK devs <>
Version: 11.19.11 (Oct2014-SP3)
Duplicates: #4050
Last updated: 2016-10-25 15:16:35 +0200
Comment 22080
Date: 2016-04-22 15:31:04 +0200
From: Richard Hughes <<richard.monetdb>>
Created attachment 394
merovingian.log
Build is Oct2014 35dd891a5d01 (2015-03-23).
I filled up the hard disk; lots of things started failing, as expected.
Upon freeing up some space, the database still refused to start: "!FATAL: Logger_new: inconsistent database, catalog does not exist"
I have attached:
By comparing (2) and (3) you can see that the sql_catalog_bid should be at 01/103.{head|tail}, but that those files do not exist. This is the immediate cause of the FATAL error above.
The SIGBUS which took down mserver5 for the first time is, I believe, fixed in more recent releases (I'm trying to get the machine upgraded, honest!).
If you believe that problems which may have caused irrecoverable corruption of this type have been fixed in more recent releases as well then feel free to close this bug with extreme prejudice. I filed this bug because I don't remember seeing any changes of this kind in release notes or checkin comments.
I don't have an core dumps for any of this stuff - they were trying to be written to the disk that was full.
Comment 22081
Date: 2016-04-22 15:31:50 +0200
From: Richard Hughes <<richard.monetdb>>
Created attachment 395
BBP.dir
Comment 22082
Date: 2016-04-22 15:32:38 +0200
From: Richard Hughes <<richard.monetdb>>
Created attachment 396
directory listing of database
Comment 24624
Date: 2016-10-25 15:16:35 +0200
From: Richard Hughes <<richard.monetdb>>
*** This bug has been marked as a duplicate of bug #4050 ***
The text was updated successfully, but these errors were encountered: