OAT HDR Issue

[PeterP]

Hi All,
Can't seem to get into iiug yet again, so posting here.

When you bring up the new group dashboard (https://www.ibm.com/developerworks/mydeveloperworks/blogs/idsteam/entry/oat_311_dashboard?lang=en_us), and have HDR servers defined it seems OAT wants to run a query with a temp table. HDR/RSS doesn't like this.

Any fixes? Seems even if you run the onmode -wf TEMPTAB_NOLOG=1 it still won't work. Am I missing something? I can't seem to set NOLOG on an HDR secondary, no matter how often I read Eric's nice write up. Presumably deprecated.
http://www.informix-zone.com/node/670

The OAT error is:
256 : Database query failed: -

Error: -229 [Informix][Informix ODBC Driver][Informix]Could not open or create a temporary file. (SQLExecute[-229] at /vobs/idsadmin/idsadmin/build/media_build/linux32/PHP/php_source/ext/pdo_informix/informix_statement.c:759)

11.70.FC7 or 12.10.FC1, OAT 3.11.

Next idea is to install an earlier OAT (or infx). Although as usual it's early/late, so if someone can point me to anything obvious other than a bug.
TIA.

Also, whilst I have expert attention, is there a specific / best way to setup trust for Informix between hosts on linux (Centos 6.4 or Debian 7.1)? Is the default RSH, SSH, PAM, .rhosts, .shosts, hosts.equiv ..?

[[e-mail]]

Temporary dbspace needs to be defined on the HDR secondary for certain queries on the dashboard.

- Set the DBSPACETEMP parameter in the onconfig file of the HDR secondary to a temporary dbspace
- Bounce the server
- If the temp dbspace has not been created, create one by running onspaces on the primary (for e.g. onspaces -c -d tempdbs -t -p /path/to/tempdbs/tempdbs_chunk1 -o 0 -s 100000 )
- Double check that the temp dbspace shows up with the 'T' flag when you run onstat -d on the HDR sec (e.g.: 45aac028 2 0x42001 2 1 2048 N TBA informix tempdbs )
-The OAT group dashboard should now show up without the error below.

Regards,
Sumanth.

Post by PeterP
Hi All,
Can't seem to get into iiug yet again, so posting here.
When you bring up the new group dashboard (https://www.ibm.com/developerworks/mydeveloperworks/blogs/idsteam/entry/oat_311_dashboard?lang=en_us), and have HDR servers defined it seems OAT wants to run a query with a temp table. HDR/RSS doesn't like this.
Any fixes? Seems even if you run the onmode -wf TEMPTAB_NOLOG=1 it still won't work. Am I missing something? I can't seem to set NOLOG on an HDR secondary, no matter how often I read Eric's nice write up. Presumably deprecated.
http://www.informix-zone.com/node/670
256 : Database query failed: -
Error: -229 [Informix][Informix ODBC Driver][Informix]Could not open or create a temporary file. (SQLExecute[-229] at /vobs/idsadmin/idsadmin/build/media_build/linux32/PHP/php_source/ext/pdo_informix/informix_statement.c:759)
11.70.FC7 or 12.10.FC1, OAT 3.11.
Next idea is to install an earlier OAT (or infx). Although as usual it's early/late, so if someone can point me to anything obvious other than a bug.
TIA.
Also, whilst I have expert attention, is there a specific / best way to setup trust for Informix between hosts on linux (Centos 6.4 or Debian 7.1)? Is the default RSH, SSH, PAM, .rhosts, .shosts, hosts.equiv ..?

[PeterP]

Excellent, working without a problem now. :)
Thank you Sumanth.

Some observations though. Even when it's all setup correctly, at first I got a 908 error. after about 60 seconds, OAT decided to forget the 908 and just work.

When I first tried DBSPACETEMP I think I might have ommited the -t option, and even when I knew the fix, I still got the error (but was now 908 not temp and I didn't check the details). So patience helps.


As for security/trust between servers for ER/HDR. It's easy but extremely varied in it's setup. Hence wondering the 'standard' best practice.Is s=6 the best way to go, or ENCRYPT_HDR or ...?
The whole linux trust is more obvious, I think one host I had to look at /etc/pam.d but otherwise /etc/hosts.equiv is fine if installing from scratch and using a normal SQLHOSTS connection. CentOS seems more for beginners than Debian. I'm writing on my website when I have time.

I found this link in the infocentre very interesting
http://pic.dhe.ibm.com/infocenter/informix/v121/topic/com.ibm.admin.doc/ids_admin_0175.htm

As for "failback" after HDR failure (without con mgr), the manuals/redbooks are a complete mess. I will likely make another post to this group, see how others handle.

[s***@gmail.com]

I suspect the 908 error would have been the cached connection status from the group summary dashboard when the HDR secondary might have been bounced. The refresh button in the right-most column of each server entry forces a refresh of the cache for that server.

Regards,
Sumanth.

Post by PeterP
Excellent, working without a problem now. :)
Thank you Sumanth.
Some observations though. Even when it's all setup correctly, at first I got a 908 error. after about 60 seconds, OAT decided to forget the 908 and just work.
When I first tried DBSPACETEMP I think I might have ommited the -t option, and even when I knew the fix, I still got the error (but was now 908 not temp and I didn't check the details). So patience helps.
As for security/trust between servers for ER/HDR. It's easy but extremely varied in it's setup. Hence wondering the 'standard' best practice.Is s=6 the best way to go, or ENCRYPT_HDR or ...?
The whole linux trust is more obvious, I think one host I had to look at /etc/pam.d but otherwise /etc/hosts.equiv is fine if installing from scratch and using a normal SQLHOSTS connection. CentOS seems more for beginners than Debian. I'm writing on my website when I have time.
I found this link in the infocentre very interesting
http://pic.dhe.ibm.com/infocenter/informix/v121/topic/com.ibm.admin.doc/ids_admin_0175.htm
As for "failback" after HDR failure (without con mgr), the manuals/redbooks are a complete mess. I will likely make another post to this group, see how others handle.