Best uptime – Y’know, one of those things you can only calculate after the machine has gone down…

Lousy power glitch. *pout*

The pertinent section of the report from apcupsd on eclipse at Neville:
XONBATT : Tue Nov 15 18:40:41 EST 2005
TONBATT : 10 seconds
CUMONBATT: 147 seconds
XOFFBATT : Wed Nov 02 20:28:45 EST 2005
LASTSTEST: Wed Nov 02 20:28:38 EST 2005

Looking at my monitoring pages, I see it was rather large… South-keys, Hog’s Back, Maitland, Island Park Dr., Bronson/Slater, Neville, Manotick(!), Kemptville(!!). I contacted Ottawa Hydro Customer Service, and the guy didn’t know anything till I asked him to check, he said there was an unusually short notice about it… typically there’s more detail. Just said “3 circuits out” and that was it. I suspect Ontario Hydro (not OPG) had a switching fault like we did back in June or July.

Anyway – the uptime record is for collocution.mWare.ca, which is the #SPORKS/mWtNet server and root-node.

This post shows that collocution hit 600 days up. I also know that I logged into it a few weeks ago and saw 667 days up, and I know it hasn’t reboot till tonight because Hybrid doesn’t auto-start.

Putting this all together… (PostgreSQL… because it’s nice for working with dates)

template1=> SELECT date_trunc(‘day’, ((now() – ‘2005/08/22 23:24’) + ‘600 days’)) AS “Collocution’s Estimated Uptime”, now()::date;

Collocution’s Estimated Uptime	now
684 days	2005-11-15
(1 row)

template1=>

And thus I declare my best ever uptime.

The previous clean-shutdown-max-uptime I believe was veisa.mWare.ath.cx with:
root@plastron:/vol/local/bigvol/_veisa/root# cat shutdownJuly1
5:47pm up 472 days, 4:14, 11 users, load average: 0.00, 0.00, 0.00