REINDEX COMMAND???

[Duy Nguyen]

I accidentally clicked on the “reindex” command from the NETWORK MONITOR.  Can someone tell me what this does?  I observed in the watch log file for the process and here’s what it says:

[cmd :cmd :INFO/0:      oru_cmd:09/20/2007 16:51:38] Receiving a command

[cmd :cmd :INFO/0:      oru_cmd:09/20/2007 16:51:38] Received command: ‘fr_quest_oru,oru_xlate reindex’

[cmd :cmd :INFO/0: fr_quest_oru:09/20/2007 16:51:38] Doing ‘reindex’ command

[cmd :cmd :INFO/0:    oru_xlate:09/20/2007 16:51:38] Doing ‘reindex’ command

[cmd :cmd :INFO/0:      oru_cmd:09/20/2007 16:51:38] Receiving a command

Hope I didn’t break anything…. 😀

Topic

[Bob Richardson]

Greetings,

You have just rebuilt your tclIndex file that contains the location of all Tcl procs in the site/tclprocs directory.  This can be dangerous if you are using Unix style version control and have files named with (say) .s for (I believe it is) CSS (unix versioning).  As I understand re-index, it is a method of reloading the paths to all of your Tcl procs in the site/tclprocs directory.  In the case of version control, you may introduce invalid names for the same proc and crash your processes.

I usually avoid doing that deliberately to avoid adding a proc to the valid list that is under development (for example).

I hope this helps you out.

[Duy Nguyen]

So I should be safe then if I understand correctly.  REINDEX is the same as the “mmktclindex” command you run from the command line?  I’m running in windows 2003 environment.

[Jim Kosloskey]

Duy,

It has been a while since I have had to consider what reindex does but this is what I recall:

Reindex rebuilds the in-memory index of the Tclprocs which are contained in the tclIndex files of all of the directories in the Autopath.

So if you were to make a new Tcl proc and do the mktclindex (so that the appropriate tclIndex file got updated) the proc still woud not be available to the Cloverleaf(R) engine because the in memory index did not include the proc.

Enter reindex. Do the reindex and the new proc now becomes known to the Cloverleaf engine – without needing to stop the engine.

Of course if you stop and start the engine, Cloverleaf’s in-memory Tcl index will get rebuilt because that is part of what happens on startup.

IMHO I don’t think you can get yourself in any trouble by what you just did as that only rebuilds the index.

Now, reload is a another kettle of fish.

Any way that is my recollection.

Jim Kosloskey

email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

[Gary Atkinson]

Just the topic I was looking for  😆 I’ll be added in my first set of code into our production site soon.  I will be adding in 3 new procs and 7 new xlates into our inbound ADT thread.  When I added these in our test site, I just bounced the process and everything was brought in.  So, this leads me to my question.  Instead of bring down the process in production (which always makes me nervous) what series of commands can I use in network monitor on the inbound thread?  I’m thinking they are “purge caches” and “re-index”?  Does “reload” need to be issued to?

Thanks in advance

Gary ❗

[Duy Nguyen]

Gary Atkinson wrote:

Just the topic I was looking for

[Michael Hertel]

Reload will refresh existing configured tclprocs.

If you tweak the proc and it has been defined to the thread when you started the thread you’re fine.

My experience has been that if you add a new proc to the thread definition then you have to bounce to get the thread to load it.

[Author]

Replies