- This topic has 7 replies, 5 voices, and was last updated 17 years ago by
.
-
Topic
-
Hello, I am trying to move my test site into a production environment. Test and production are on two different servers, not two sites under the same Cloverleaf service.
My deployment strategy was to copy the translations, tclprocs and netconfig from the test box to the production box. I started one of the processes in the site but when I tried to start the monitor daemon I got this …
[msi :msi :ERR /0: hcimonitord:05/09/2007 12:41:02] msiShmemAttach: Can’t create shmem file ‘C:/quovadx/qdx5.4.1/integrator/bigthumb/exec/monitorShmemFile’: No error
[msi :msi :ERR /0: hcimonitord:05/09/2007 12:41:02] msiGlobalInit: Couldn’t attach memory region
[prod:prod:ERR /0: hcimonitord:05/09/2007 12:41:02] Initialization of global shared memory failed
… and the monitor panics.
If I run the monitor first and then try to start a process, the same thing happens to the hciengine, it panics with the same message. It’s as if the monitorShmemFile can’t be accessed by two programs at the same time.
I tried deleting the monitorShmemFile but that didn’t help. I also tried a number of recipes that I found in the forum, things like hcimsiutil, hciinitdb and hcisiteclean.
I’ve even checked the permissions on the file and they are the same on both the test box and the production box. BTW, we are running on windows server 2003.
What should I look at next?
Thanks for your help.
Dennis
-
Replies
-
-
Maybe you already tried all of these, but this is what I was told to do bring the site down and then back-up: 1.) hcieginestop -p
2.) hcisitectl -K
3.) hciss -kh
4.) hcimsiutil -R
5.) delete vista.taf from databases directory
6.) delete monitorShemFile from exec directory
7.) hcidbinit -i
8.) hcisitectl -S
9.) hciss -sh
10.) hcienginerun -p
-
-
-
Quote:
My deployment strategy was to copy the translations, tclprocs and netconfig from the test box to the production box.
It sounds like you copied the exec directory from the test site or files from the exec directory.
This is a no-no.
You should delete and reinitialize your site, then copy what you want to your site but leave the “exec” directory alone.
The site should be created with the siteinit tool from your workstation.
-
-
I encountered the same problem yesterday. I could start/stop existing sites, but when I initialize a new site and try starting up threads, it would result in that same error. Solution that worked for me:
If I rename/delete the monitorShmemFile to something else, then it works fine the next time I try to start/stop the threads.
-
FYI – sometimes gremlins do bad things and the next thing you see is an evil file in the sitePorto called $HCIROOT/siteProto/exec/monitorShmemFile
and that will manifest symtpoms similar to what I see described in this post until you delete it; not to mention shared memory problems that are so tough to get rid of you may have to delete any newly created sites and start over to completly remedy.
There are so many ways this can happen I want go into all of them but anything that does a regular sweep of all the sites (especially if it calls msiAttach without checking that monitorShmemFile already exists) has the potential of hitting the siteProto adversly so take care.
Perhaps the simplist mistake would be to unintentionally launch a netmonitor while in siteProto.
If you think not then do
ls -l $HCIROOT/siteProto/exec/monitorShmemFile
and see if you don’t magically have a hit.
I have a cron entry run daily safe keeping type task and one of them is as follows
rm -f $HCIROOT/siteProto/exec/monitorShmemFileEven if this never happens it is another safe guard that does not hurt anything and could end up being of help.
Russ Ross
RussRoss318@gmail.com
-
- The forum ‘Cloverleaf’ is closed to new topics and replies.
