Jump to content

Search the Community

Showing results for tags 'hang'.

  • Search By Tags

    Type tags separated by commas.
  • Search By Author

Content Type


  • Announcements, News and Resources
    • Latest News
  • Windows Products-Retrospect
    • Professional
    • Server, SBS and Multi Server
    • Device and Hardware Compatibility-Windows
    • Exchange Server Add-On Support
    • SQL Server Agent
    • Linux, Unix and Netware Clients
    • Express for Windows
    • Product Suggestions-Windows
  • Mac OS X Products-Retrospect
    • Retrospect 9 or higher for Macintosh
    • Retrospect 8 For Macintosh
    • Retrospect 6: Desktop, Workgroup and Server for Mac OS X
    • Device and Hardware Compatibility-Mac OS X
    • Linux Clients
    • Product Suggestions-Mac OS X
  • Macintosh OS 9 and Earlier-Retrospect
    • Express, Desktop, Workgroup and Server for Pre-OS X
    • Device and Hardware Compatibility Pre OS X
  • General Discussion-Retrospect
    • Networking and Clients
    • Strategy, Scripts and General Use
    • Retrospect iPhone App
  • Retrospect 8.x for Mac
  • Retrospect 6.1 for Mac
  • Retrospect 7.7 for Windows
  • Retrospect 7.6 for Windows
  • Retrospect Express
  • General Discussion


There are no results to display.

Find results in...

Find results that contain...

Date Created

  • Start


Last Updated

  • Start


Filter by number of...


  • Start





Website URL







Found 6 results

  1. An old issue has been "revived": - During a cleanup task the engine does not free its private memory running into a out of application memory issue Initial setup (already set at version 10 and run at least for 60 times successfully before upgrading to 11.0.0-194): - weekly scheduled cleanup of all mediasets - mediaset settings vary from "retrospect policy" to "last 10 times" depending on the mediaset content and type The script runs each friday night at 8 pm for about 12 hours. This works fine until the upgrade to 11.0.0 The first friday after upgrading to 11.0.0 the task runs as well, but leaving my system unusable with a memory issue "out of application memory". Holding the engine and freeing 500 MB of memory by closing some open applications, shows, that the engine consumed 24 GB of private memory and additionally 23.9 GB of kernel memory. Stopping the engine by killing the process 'kill <process-id>' freed all memory instantaneously. While trying to reproduce the issue, I manually started a cleaning/grooming task: 1.) taking memory footprint of retroengine (0.034 GB physical private memory / 2.61 GB virtual memory) 2.) starting cleanup task 3.) waiting for 10th+ collection stage 4.) taking another memory footprint of retroengine (22.3 GB physical private memory) 5.) stopping the task using the retrospect console, leaves the engine with still allocated 22.3 GB private memory Please see enclosed screenshots for further information of steps 2 to 5. Possibly there are two issues with retroengine 11.0.0: a.) excessive memory usage whilst cleaning b.) not freeing memory between stages Unfortunately I cannot go back to version 10 as I already migrated my mediasets! My configuration is: MacPro 12 core 48 GB RAM 2 TB RAID HDD 15 TB NAS 12 TB FW attached Backup Storage Ticket 27403 - Screenshots .pdf
  2. I hit a hang at some point since the 26th of January. I saw the engine busy doing a scripted verify of all my sets on disk yesterday (It appeared to be running, from the flashing lights on the external disk(s).) Unfortunately, it was still "running" today, with the Retro engine at 100% CPU (one CPU @ 100%), and no lights flashing. There were two scripts showing "busy" in the Retro console. I tried to stop the two activities, and tried to stop the engine the "normal" way through the CP. I had to force quit the engine to free it. I snapped the screen with the two activities, and also the activities after engine restart. Note that the two activities that were "done" have vanished - no evidence that they happened. I include screen snaps of the two activities with both summary and log, and also a screen snap of the console after recovery. THe sample, as taken by the Activity Monitor while the engine was hung. This is a bug. Retro was "stuck" for almost a week. (Retro engine 10.5.145, Mac OS X 10.8.5 - Console retro 10.5.145, on Mac OS X 10.6.8) Sample_RetroEngine.txt
  3. Reporting the news. All retro 10.5.1 (145). Console running on mini, 10.6.8. Engine on 10.8.5 on newer mini. 16 GB mem. I had the console up for a while. I clicked on "refresh" for one of my clients, and it seemed to take a while. I put the console in the background and forgot about it. A good hour later, I brought the console to the front, and it was *still* spinning, waiting for access to the client. I clicked onthe little "x" to tell it to quit, and then clicked on the client, and the "refresh" again. I don't remember exactly what I did at that point, but the console stayed up, and the engine appeared to crash. The display collapsed to the icon for the engine/server, and it said "connecting". I tried to restart the console, and it did not help. It also refused to connect the the enfine. I moved over to the server, and being aware of the problems with retro crashing and messing up its prefs, I decided I would see if there was a crash log, and maybe do something about the prefs *before* i restarted. On the server, the engine was still running, but refusing all console connections. As I recall, after I restarted the engine, the console still would not connect. Once the engine was up, restarting the console worked. I went to the control panel on the server, and noticed that "instant scan" was enabled (I KNOW I disabled it and left it disabled) I unlocked it and pressed the stop button. The engine stopped. I then went to /Library/Applications Supprt/Retrospect, and noticed that config80.dat was 14.7 MB, and config80.bak was 1.7 MB. This looked sub-optimal. I don't want one of those huge config80.dat files. I moved config80.dat to config80.dat.bad, and then I copied config80.bak -> config80.dat. I also copied config80.bak to config80.bak.bak (just in case - not certain retro is actually backing up its pref files) I restarted the engine, and all is well EXCEPT that the "activities" prior to today are all gone. I can see the activity in the log, and I can see the passt backups, but the stuff in the "activity window is gone. Small price to pay. Why did it crash? Why did it crash without actually crashing? Did the "non-crash" cause the "large" config file? Why did the crash "re-enable" the insta-scan? What state in the console would prevent it from connecting to the engine?
  4. When I do certain operations, like try to mount an nfs mount point from the "add sources" pane in the console, the machine that is running the engine can put up a dialog on the screen, and the operation on the machine running the console hangs. The console can not proceed until the dialog is dismissed. The most recent case of this is mounting NFS from that "add source" dialog, but I know I have seen this before. It is not obvious that you have to have access to the "screen" of the engine-local system to operate the remote console, but in some cases, you do. A remote console should never put up a dialog on the engine's machine that the user of the console can't see.
  5. I have changed my local IPs, so I had to re-add the engine(s) to the instance of the Retro 10.0.1 (105) console running on my laptop. i added the "local" engine first by deleting the two engines I had in the left sidebar. I added the local engine with the new IP and passwd. I added the "tunnel" engine(s) and tested one, then "killed" the tunnel, causing the second engine to disconnect from the console. At that point, the panels in the left sidebar of the console for the local machine (virtue) disappeared. I waited a while, and it did not come back. I noticed that the CPU was in a tight loop on the laptop. I did a sample from Activity Monitor, and have included a snap of the console window, and the sample text. You should be able to figure out where it's stuck. Enjoy. Sample of Retrospect.txt
  6. I had this problem for the past week and figured out what the problem was so thought I'd put it out there. The problem (kind of) wasn't Retrospect in this case. The Retrospect server (Mac, Retrospect 8.2) kept dying backing up a particular script on one of our file servers (CentOS 5.6). A reboot fixed this for a day but then the problem reoccurred and frequent reboots of the file server are impractical. The symptom: I narrowed the problem down to one script which backed up the root partition. Whenever this ran, Retrospect hung on the script and after stopping it, all other jobs on the server failed. The cause: I logged into the server to try to figure out the problem and ran "ls /" which worked as usual. I then ran "ls -l /" which hung indefinitely. I ran "strace ls -l /" which told me that the hang was happening when accessing a directory mounted off the root which is a hard nfs mount to an external drive. This is the problem with hard nfs mounts - they can hang indefinitely when they fail. Once I fixed the mounting issue, everything was gravy. So this wasn't entirely Retrospect's fault. However, once I stopped the hanging script, the server should have released the client connection (I ran "rcl status" and the connection was not being released). Hopefully this helps someone else keep their hair!
  • Create New...