Month: December 2016

Splitting a btrfs 4.1x root partition with debian live system gparted btrfs tools 3.17-x


This is a short picture log of doing a btrfs 4.1x root partition split on a (down) oracle linux 7.2 using a debian live system applying gparted based on btrfs tools 3.17-x. Lot’s of names and version codecs, right? But this is what matters. The important message is : it works using this flavours.
Actually, running oracle or redhat linux as the live system may have been much more appropriate concerning compatibility reasons. The odd things is, no redhat-based (enterprise) linux system features gparted. Only fedora does, sourcing the epel-repository but not having kinf of a live system release as debian.

(more…)

Advertisements

The look of : transaction deadlock due to missing foreign key index


This about transaction deadlocks resulting from an index missing on foreign key colum(s). Actually, the post will not dive into the odds resulting from this failure but show how to detect this dedicated reason for an “ORA-00060: deadlock detected while waiting for resource” message. In particular, what to look out for in the trace file and how to find the table objects (instead of rows) being deadlocked against each other. The best resource on the web tackling this subject is of course (Reading deadlock trace files) from Tom Kyte, which also guided me along the way of a “no row deadlock”, see below.
Ok, on first place, some tool, Enterprise Manager for example, or whatever will notify you about entries in the alert.log, reading roughly the like as “Sun May 29 02:51:07 2016 ORA-00060: Deadlock detected. More info in file H:\ORACLE\diag\rdbms\got\got\trace\got_ora_4536.trc“. Inspecting the trace file, you should focus your attention on the first 70 lines, in most cases neglegting the trailing process information dumps. I got this, anything noteworthy will be highlighted and discussed below:

(more…)