How to stop/remove stalled ArcGIS Server cache processes

7072
7
Jump to solution
07-22-2014 07:03 PM
MikeLouwrens
Occasional Contributor III

I have a server running ArcGIS Server 10.1.  I've been trying to cache a basemap service, however it has been stalling and I can't cancel or remove the stalled processes, and can't get it to continue by creating new processes.  How do I cancel and remove those stalled processes?

thanks,
Mike.

Tags (2)
0 Kudos
1 Solution

Accepted Solutions
DivyamGulati
New Contributor II

Hello Mike,

In case you want to stick to 10.1, I would suggest you to upgrade to 10.1 SP1 Quality Improvement Patch. Many bugs and other issues related to caching have been addressed with this patch.

http://support.esri.com/en/downloads/patches-servicepacks/view/productid/160/metaid/2064

Also, a good approach in these cases is to examine status.gdb created in the cache directory. You can create a copy of status.gdb on your local machine and manually examine 'task status' and 'job status' feature classes by pulling them in ArcMap. These feature classes may convey valuable information (possibly an error) about the extent and level the cache creation is failing. You can then check for the elements in that extent which might be stalling the caching process.

Stalled processes can perhaps be killed by manually removing the processes from the arcgisjobs > caching service directory.  You need to stop the caching service before manually deleting theses jobs.

-Best from Divyam.

View solution in original post

7 Replies
DavidColey
Frequent Contributor

Hi Mike-

Could you provide a bit more info on 'stalled processes'?   Can you supply an error message? 

Thanks

David

0 Kudos
MikeLouwrens
Occasional Contributor III

Hi David, thanks for your reply.

I have no error, they're just not doing anything.  When I check the cache status it says something like "Processing 43%" and it may have been there for a week.  It got from 0 to 43 in a couple hours.

I can't cancel the process, when I click cancel it doesn't do anything.  There are no warnings or errors or anything to tell me it's not working.

Cheers,

Mike.

0 Kudos
DavidColey
Frequent Contributor

Sure.  Um, well as you know caching is very I/O intensive for enterprise databases.  You can try a couple things: First, if you must use your SDE as your source layers, try backing out some of your more intensive layers such as parcels, annotation and the like and run some high-level tiles and check your performance.  Then keep adding in layers and labelling combinations until you can find the offending layer.  Or, if you can, go ahead and copy your cache layers to a fdgb that resides in your data-store, resource your mxd to those layers, re-publish your service and then re-cache.

For us, I found that for our county basemap, with scales down to 1:600 (for 600 mi2), cached at about 4 hours when running a full 14 caching instances on our gp server.  Again, that was when sourced to our enterprise SQL SDE. When I replaced the source to fgdb in our data-store, that time decreased to 20 minutes.

Hope this helps

David

MikeLouwrens
Occasional Contributor III

David Coley wrote:

Sure.  Um, well as you know caching is very I/O intensive for enterprise databases. 

Well that's the thing, CPU is sitting around 3%, I don't think it's doing anything.

Cheers,

Mike.

0 Kudos
ShaunWeston1
New Contributor III

Has the caching actually stopped? Is the CachingTools GP service running, try restarting that. Have you had a look at the "Manage Map Server Cache Status" GP tool? That tool can remove the cache status information (what tiles have been built and how far through) and rebuilt it.

MikeLouwrens
Occasional Contributor III

Shaun Weston wrote:

Has the caching actually stopped? Is the CachingTools GP service running, try restarting that. Have you had a look at the "Manage Map Server Cache Status" GP tool? That tool can remove the cache status information (what tiles have been built and how far through) and rebuilt it.

I tried running the "Manage Map Server Cache Status" tool.  Started it at 9:46AM on Friday, I killed it this morning about 7:30AM Monday - all it had done was clear my status, so now I see nothing in the "View Cache Status..." window.  It doesn't even list the cache levels in there anymore?!  Now when I run this tool again it fails, telling me that "The workspace is not connected"

I restarted the Caching Tools service, the "View Cache Status" window now shows that Tile generation is not in progress...

I will try to delete all tiles and start again... I'll also try an update either to 10.1 SP1 or maybe even to 10.2.2

Cheers,

Mike.

0 Kudos
DivyamGulati
New Contributor II

Hello Mike,

In case you want to stick to 10.1, I would suggest you to upgrade to 10.1 SP1 Quality Improvement Patch. Many bugs and other issues related to caching have been addressed with this patch.

http://support.esri.com/en/downloads/patches-servicepacks/view/productid/160/metaid/2064

Also, a good approach in these cases is to examine status.gdb created in the cache directory. You can create a copy of status.gdb on your local machine and manually examine 'task status' and 'job status' feature classes by pulling them in ArcMap. These feature classes may convey valuable information (possibly an error) about the extent and level the cache creation is failing. You can then check for the elements in that extent which might be stalling the caching process.

Stalled processes can perhaps be killed by manually removing the processes from the arcgisjobs > caching service directory.  You need to stop the caching service before manually deleting theses jobs.

-Best from Divyam.