So I was working on a project using Creo 2.0 connected to Windchill 10.2 when all the sudden there is a problem with check-in of my part. I try to investigate - I click on the red stop sign on the status bar to examine the event manager - up comes a blank window.
I try the "workspace" link - up pops up an error message with a Status Code 501 - no access to this object or object does not exist. Uh oh, that 2nd part doesn't sound too good.
I check in a stand-alone browser - things don't look good: many of my workspaces have disappeared: they are no longer listed on the "My Workspaces" section in my home page.
There are still links to these objects in the "Recently Visited" lists, but attempting to view them throws the same Code 501 message.
I ask other users if they are having issues - nope.
Logging in as the wcadmin, I can see that the workspaces for my regular user are still there. But from there, the only thing a wcadmin can do with them is to delete them...
So, looking at the methodserver logs, I see ERRORS all over the place saying my regular user doesn't have permissions to access the context(s)... What the heck? No one was messing with access rights, all users still belong to the same groups that have the correct permissions... It's a total mystery to me why I am all the sudden locked out of some product contexts.
So I get PTC tech support and they recommend I go and grant my user access to these contexts through the policy administration tool. Well, after giving my user the Read rights to some PDMLink object of a particular context, "My Workspaces" section is now fuller. But when I try to access the common-space folders of this context, the same 501 error comes up. So the tech support say we need to add more Policies that grant my user read access for the Cabinet type of an object... I am starting to panic because this solution doesn't seem to be getting at the root of the problem - am I going to have to patch the policies for every context I was involved in before?
So I ask - can we try to reboot the server first?
And after doing so, the problem goes away.
Well, there is another thing - I used to have bookmarks in the "Hot Links" of the "My Notebook" section in my homepage. Now they're gone.
So questions to you Windchill gurus:
- has anyone seen this kind of flakiness before?
- does Windchill automatically get rid of the bookmarks that point to contexts to which you don't have access?
- the server has been up for 17 days before restarting it today. Is this too long of a time? We do have a daily backup going - but I wonder if the windchill server should be shut down during the backup process and then restarted after the vaults are backed up...
Can't really answer the first two questions...
However, we shut down the Windchill services every night during our daily maintenance routine. The computer doesn't get rebooted very often, but those services get shut down and restarted every night.
Don't really have a whole lot of flakiness that I have been made aware of.
I have heard it said that you should practice not until you get it right, but until you can't get it wrong. Maybe software needs to take the same approach.
CAD / PLM Systems Manager
I reboot my Windchill servers every morning at 5AM just to be sure we don't encounter these sorts of errors.
How often do you restart Windchill and/or reboot your servers?
"An ounce of prevention is worth a pound of cure." - Ben Franklin
Windchill 10 m040
Creo 2 m220
I would do some digging through your logs before this started occurring and see if you can find anything that might indicate why. It almost seems like something died and automatically restarted, but did start up correctly.
As far as server backup and restarts, we only restart when installing a critical patch set (or making some other property change.) Typically this happens every month or two. Our backup software takes snapshots of both the application server and the database server while the system is still running. I've seen no performance difference (at least here) between a freshly restarted system and one that's been running for two months straight. (We have multiple locations with users around the world, so daily restarts would cause issues for us.)
Ah, so you have a backup system where the data is constantly backed up? I am wondering how it is all synchronized up.
We do a daily incremental "save-to-tape" function. Shortly before this happens, there is a script that copies some key files to a designated folder, one of them being the oracle database dump file. I am at this point trying to find out if Windchill should at least be stopped during the execution of this script, so that no user can interact with the database (by renaming objects, or checking out , etc...), as its trying to do its data dump.
Anyway, at least I found out that your "Hot Links" bookmarks do get "purged" automatically if they start to point to contexts to which you no longer have access.
Would you be interested in a quick assessment of your system to point out any opportunities for improvement? Please let me know and I'd be happy to get some information to you.