Hello everyone! This is quite an involved problem but I'll try to keep it
short.
We're running Informix Online Engine 5.08 on an AlphaServer 2100A.
Yesterday we had to replace a disk. So we brought down the chunks on that
disk, replaced the disk, rewrote the disklabel, and then brought the
chunks on that disk into recovery.
In midst of rebuilding these chunks on this one database instance (our
test instance), activity on the production instance (not involved
disk-wise) came to a halt (angering a lot of people, not fun in a
political environment either)!!!
After a call to Informix support, they said 'it must be a system issue'
8-p to which I concur. The only thing I can put my finger on is that the
system max-ed out the available IPC resources.
Our kernel configuration, put together by a company no longer willing to
help us put this in the configuration file:
semmns 512
semmni 64
semmsl 128
shmmax 8388608
shmmni 128
shmseg 64
I know what these variables are, but I don't know what the interpretation
of them is.
What units is the shared memory size in?
Can I solve our problem by increasing some of these parameters?? If I do,
what impact on system resources/performance does it have??
Suddenly feeling like I'm out of my league...
Karl
--
Karl Marble
Unix Administrator & Webmaster
Information Services
City of Worcester Massachusetts
mailto:marblek_at_ci.worcester.ma.us
Received on Wed Mar 12 1997 - 22:35:08 NZDT