Self-healing DEHS
Saturday, 1. November 2008, 19:26:17
If someone by any chance has noticed in
DEHS' stats page
that there was a bump in the number of failing watch files and is
wondering why it hasn't gone down like it used to, here's the answer:
I'm testing how DEHS' can heal itself, and how long it takes.
In the past when such kind of bumps occurred I used to start a special
DEHS run which retries all the failed watch files, but since doing
that is probably the reason why some of the normal runs process many
more watch files and take longer to finish I decided to let DEHS
automatically recover.
According to my plan it should take from 6 to 7 days to completely
recover from the last mass-failure which was caused by sourceforge's
garr mirror going down abruptly. And I hope this information will be
of some use to design a proper self-healing/recovery mechanism for
DEHS2 (or should it be DEHS-ng? I don't know yet, feel free to propose
names)
DEHS' stats page
that there was a bump in the number of failing watch files and is
wondering why it hasn't gone down like it used to, here's the answer:
I'm testing how DEHS' can heal itself, and how long it takes.
In the past when such kind of bumps occurred I used to start a special
DEHS run which retries all the failed watch files, but since doing
that is probably the reason why some of the normal runs process many
more watch files and take longer to finish I decided to let DEHS
automatically recover.
According to my plan it should take from 6 to 7 days to completely
recover from the last mass-failure which was caused by sourceforge's
garr mirror going down abruptly. And I hope this information will be
of some use to design a proper self-healing/recovery mechanism for
DEHS2 (or should it be DEHS-ng? I don't know yet, feel free to propose
names)
WIDGETIZE
Anonymous # 2. November 2008, 04:35
Just call it DEHS and announce the improvements, thanks.