H Rrefresh Failing on prod State

book

Article ID: 208479

calendar_today

Updated On:

Products

CA Harvest Software Change Manager

Issue/Introduction

HRrefresh Failing on prod State

For one state in one project, HRefresh completes the "delete-check" step and lists 19 files that don't need to be deleted, then throws:

E03060086: Checking file synchronization status failed.
I00060080: Check out summary: Total: 365 ; Success: 0 ; Failed: 0 ; Not Processed: 365 .
Checkout execution has failed.

- Other HRefresh jobs targeting sibling folders on the same agent machine don't show any similar issue.
- Took the hsync command from the HRefresh log file, turned that into an "hco -sy" command and that worked without any problem.
- Tried deleting all the sig files - did not resolve the problem.
- Tried changing from "Full" mode to "Active" mode - did not resolve the problem.
- This one project and state fails in the same place each time, all other jobs (as far as I know) targeting the same agent and same general folder area work fine.
- Confirmed agent and broker are the same version of SCM.

Environment

Release : 12.6

Component : CA Harvest Software Change Manager

Resolution

We found 2 workarounds for the problem.  One was to use the "hco" command instead of "hrefresh" to deploy the files,  The other was to delete all the files (not just the .harvest.sig files) in the client path folder before executing hrefresh.  Since his version of Harvest is no longer supported, we were a bit limited.  Our options are:

  1. We could continue to try and diagnose the problem.  The fact that deleting all the files before hrefresh allows it to succeed seems to indicate a problem with one or more of the files in that folder.  The fact that hrefresh for other projects, and other states of this project always succeed would seem to indicate something different about the set of files in the Production state.  We could use BeyondCompare or WinMerge to do a folder-to-folder comparison between the deployment folder for the Prod state with the deployment folder for one of the other states to try and find what might be different.
  2. We could change the UDP properties in this one state to use the hco command instead of hrefresh, since hco can successfully check out the files.
  3. We could add a pre-processing step before hrefresh in this one state to delete all the files first.  Here's how to do this:

In your hrefresh.cfg file, for the project and state where we’re having the problem, add to the end of the line:

, , PreCmds: { rm -Rf /client/path/name }
(replace “/client/path/name” with the target folder for the hrefresh to update)

Here’s the example from my test box:

SampleProject,Development,Check Out for Browse,\SampleRepository,/home/cascm/dev,Full,olnx77,6000,harvest.dfo,agent.dfo,,PreCmds: { rm -Rf /home/cascm/dev }