Known Issues upgrading from 5.5 to 5.7

[Chris Wethey]

Looking for heads up information prior to starting the process of upgrading our IM 5.5 to 5.7 – any gotcha’s that any of you have had with the upgrade.

Any information is greatly appreciated.

Thanks

Chris Wethey

Topic

[Lawrence Williams]

We went from 5.4 to 5.7 and here are a couple ‘gotchas’ that I remember from the top my head….

–  We had a httppost interface and since 5.7 uses cURL we had some issues with the interface, we ended up working around it with some extra TCL code

–  We haven’t installed any REV patches yet, but I am trying to get REV2 installed….the biggest reason being the SMAT searches, with 5.7 you have to build a ‘search definition’ any time you want to find a particular string within the SMAT messages and it is a pain when you only want to do a quick search

–  One that I am fighting with now, can’t say whether it is 5.7 related or not is the LogHistory.  In a nutshell, it’s not consistent….when I enter that I want the logfile cycled when it reaches 10,000 KB sometimes it works and sometimes it doesn’t and it has caused processes to die with no descriptive error to troubleshoot with.

[Nate Kruse]

We just upgraded from 5.5 to 5.7Rev2 for our production environment.

[Amy Gray]

QDX5.7 issues alot more warning for XML schemas then before.    Also, we have several TCP IP threads getting SEPCHARS errors even though they are not XML

[Jerry Hayes]

We just did a 5.5 to 5.7 upgrade.  Issues:

1. Default date precision is ON in 5.5 and OFF in 5.7.  Our data compare process required us to write a utility to add precision ON in 5.7 translates (that didn’t already have precision set), and fix a few TCL files with the convert_date procedure (which uses the same precision internally).

2. The new RAIMA configuration wouldn’t let us continue to count recovery/error records more quickly with the keydump utility.  We had to switch these to use hcidbdump.

We had a variety of other small issues that depend on specific implementation details, but those were the largest factors.

As a side note, we expected the RAIMA to be corruption free, but were still able to induce corruption in power-off setting, (not using HA). This was able to be triggered even in VMWare (which we use for our failover/etc.).

Hope this helps, feel free to PM me if you need any more discussion. Our implementation covers about 250+ threads.

Jerry

[Amy Gray]

Hello,   we tried installing 5.7 on AIX and although the interfaces worked fine, or appeared to, we were unable to get the host server running.   No GUIs could be loaded and we eventually reverted back to 5.5.    Tech support checked logs etc but were never able to determine the problem.   We will be installing 5.7 on a new AIX server shortly and hope the problem does not resurface.

[Steve MacDonald]

We are upgrading from 5.5r1 to 5.8 on AIX 5.3 Tech Level 9.  Some of the issues we have found so far……..

1)  When Bulk Copy or Path Copy is not used the outbound message will differ from 5.5 to 5.8.  In 5.5, the EVN is not optional so the first field is populated with Variant Message Type (eg. A01).  In 5.8 you will just receive the segment ID of EVN with no fields populated.

2) Using a temp variable in the outbound field pathname will not work in 5.8 as it did in 5.5.  The Temp variable will need to be changed to $%f#.

3) When running version 5.5 and 5.8 on the same AIX server, bouncing the 5.8 host server requires an ‘unset DISPLAY’ command before starting it back up.  Still checking on this one   😕

4) Outbound threads left with the default Auto-Reconnect setting of Reopen the connection every 10 seconds were filling up the logs fast (even with EO set to DISABLE_ALL).  I changed these to 7200 on our test connections.

5)  Global Monitor 4.0 ‘Notes’ do not show what was placed in the NetConfig Notes Tab.  I think the fix on this is coming out in the first rev patch though…

[Author]

Replies