CLOVERLEAF 5.7 rev2 and CLOVERLEAF 5.8 rev5

[mike brown]

Input please :

We are thinking about CL5.8R5 would the change to difficult going from CL5.7R2.

i have done other upgrades to this point, but was wondering what was in store if we went to CL5.8R5

thx mike

Topic

[Bob Richardson]

Greetings,

We have the exact scenario that you describe for an upgrade path here.

If you are going to run only ONE root then that helps enormously.

(see the AIX shared memory warnings – that is, if you are on AIX).

Read ALL the Release Notes for Every patch level.

Some BIG items:

(1) HL7 V2 variants have been fixed for those groups which should be defined as repeating groups to make them unique (like an A17 with two PID/PV1 groups at the same level).   You must make a decision to convert your existing variants (and now retest ALL of them) OR bypass conversion.

(2) 5.8.5 really complains now (and data loss is possible) if your HL7 V2 variants and Xlate paths do not match or you run into (at least so far) those ROL groups that are out of whack.

(3) Fileset/local protocols behave differently – less efficient:  we found that there was a 35% increase in latency for message processing.  Rumor has it that the file is opened/read/closed for each message rather than opening/reading and closing at shutdown.  Unconfirmed but would fit the 35% lag time in processing.

(4) Various GUI misbehaves – but (!) we like the copy/paste between Xlates and NetConfigs!!

(5) New keys added to Table objects for encoding values – not portable between 5.7 and 5.8.

(6) SMAT is changed to include message metadata in the “.msg” files now.

My guess is that this supports the new Global Monitor 5.5 ability to index and display SMAT messages.   Any custom logic that extracts messages will have to be reworked.

And.. other items that we found significant to us but these appear to be universal.

Good luck!

[mike brown]

Thanks Bob

I will research it to the tee…

mike

[Tiffany Bohall]

Hi Mike,

Hope all is well with you! Where is your sense of humor when I need it? 😆  

I felt compelled to respond because we just upgraded our test environment to v5.8.5 and are having significant issues with SMAT cycling.  See CloverTech thread titled “CL 5.8 rev 5 patch issue.”  I finally had to open a CASE with Lawson because I was not getting very far with recommendations here.

Take Care,

Tiffany Bohall

[mike brown]

Thanks Tiffany

Good to know I will address the issue you mentioned.

mike

[Alice Kazin]

The table issue is not an issue:  If you save the table in 5.8.5, it adds the Metadata line.

[Alice Kazin]

I meant the encoding line

[Alice Kazin]

There is a script which is provided that converts new Smat file to old Smat file format called hcismatconvert:  For example:

hcismatconvert -b soft_in -o alice.soft

where soft_in is the name of the current smat file and

alice.soft is my output file

You will now have a Smat file in old format “alice.soft.msg” and alice.soft.idx”.

[Mark Perschbacher]

I am about finished evaluating the upgrade of the same versions you are looking at.  The only big thing I have found is that with 5.7, the tcl gdbm function has been replaced by sqlite.  In the long run this is probably a good move, but I’m working with Lawson to develop a solution for a proc I have that is not working after the conversion to sqlite.

[Earl Zhang]

We upgraded to 5.8.4 from 5.4 last year. One thing we experienced is the SMAT files are much bigger, I guess because of all the index information. We ran out of archive space just 2 months after upgrade, so the archive disk space needs to be revisited.

[Author]

Replies