Archives of the TeradataForum
Message Posted: Thu, 28 Feb 2002 @ 14:58:35 GMT
| Subj: || || Re: Size of Crashdumps |
| From: || || Christopher Platt |
We use 10 gb on our system... we have a six node 5200... I consider this to be empty space (which it is most of the time)... so I
consider it to be part of spool space. I usually allow 25-30 % of my total space on the box for spool space. I take crash dump space out
of this 25-30%. It used to be after three crashdumps or restarts (most restarts cause a dump) the database would stay down... I have found
dumps on our box to be 2-3 gb each... (depends on the cause) I have found 10gb to hold three dumps... if there is not enough room in crash
dumps... your dump stays in Unix filesystems... And I believe it issues messages either into /etc/.osm or /var/adm/streams logs telling you
it cannot take a dump... (i.e. you are constipated !). We have our own home grown program to monitor the filesystems ... which has alerted
us to constipation. We had an issue where one of the upgrades resized crashdumps and we did not notice it that day... our box restarted and
we did not have enough room for a dump...
oh well ....hope this helps....