"Retrospect Client is off", but using 100% CPU

[Lennart_T]

Retropect Client 12.5.0 (111), Mac OS X 10.9.5, MacBook Pro 15" 2011.

 

From time to time, maybe once a month, I get the message the "Retrospect Client is off". But still I hear the fans running, which I normally don't hear.

Looking in the Activity Monitor, I see the retroclient runs at 100% CPU.

 

A restart (of the Mac) usually corrects the problem.

 

But why is it happening and can I prevent it?

 

[Lennart_T]

Now the client is on, but uses 100% CPU anyway. Why?

 

[Mayoff]

It sounds like the retroclient (or retropds) process crashed. Check the OS X Console Utility for crash reporter logs from the client.

[Lennart_T]

Thanks for answering. This is what I found:

Date/Time:       2016-06-08 21:41:15 +0200
OS Version:      10.9.5 (Build 13F1808)
Architecture:    x86_64
Report Version:  18

Command:         retroclient
Path:            /Library/PreferencePanes/Retrospect Client.prefPane/Contents/MacOS/retroclient
Version:         ??? (???)
Parent:          launchd [1]

PID:             96
Event:           cpu usage (microstackshots only)
Thread:          0x37a (88% cpu over 103 seconds)
Duration:        103.00s
Steps:           103

Hardware model:  MacBookPro8,2
Active cpus:     8
Fan speed:       3605 rpm


Powerstats for:  retroclient [96] thread 0x37a
Start time:      2016-06-08 21:41:26 +0200
End time:        2016-06-08 21:42:51 +0200
Parent:          launchd
Microstackshots: 33 samples (32%)
Primary state:   33 samples Non-Frontmost App, Non-Background Priority, User mode
User Activity:   0 samples Idle, 33 samples Active
Power Source:    0 samples on Battery, 33 samples on AC
  33 thread_start + 13 (libsystem_pthread.dylib) [0x7fff8d468fc9]
    33 _pthread_start + 137 (libsystem_pthread.dylib) [0x7fff8d46472a]
      33 _pthread_body + 138 (libsystem_pthread.dylib) [0x7fff8d464899]
        33 threadSpawner + 64 (retroclient) [0x10002596c]
          33 netEnumInterfaces + 240 (retroclient) [0x100009ac5]
            33 NetCheckNewInterfaces + 1084 (retroclient) [0x1000016f3]
              16 SThreadYield + 5 (retroclient) [0x100025c33]
              10 IPNSStart + 436 (retroclient) [0x100008a80]
              4  IPNSStart + 448 (retroclient) [0x100008a8c]
              3  SEventSignal + 10 (retroclient) [0x100025c2e]

  Binary Images:
         0x100000000 -        0x100048ff7  retroclient <FA76A00B-7644-3233-8867-22060E9898B4> /Library/PreferencePanes/Retrospect Client.prefPane/Contents/MacOS/retroclient
      0x7fff8d463000 -     0x7fff8d46aff7  libsystem_pthread.dylib (53.1.4) <AB498556-B555-310E-9041-F67EC9E00E2C> /usr/lib/system/libsystem_pthread.dylib

[eppinizer]

Hi Lennart,

 

Is it just this one client that you notice the behavior with? Have you noticed any recent errors or interrupted backups of the client?

 

It looks like we crashed when we attempted to enumerate your network interfaces.. 

 

Do you have any Proxy/VPNs that you use infrequently?

 

Any interesting network configuration, or virtual interfaces?

 

If you check your retroclient logs in /var/log, are there any interesting entries around the time of the crash? (2016-06-08 21:41:26)

[Lennart_T]

Hello.

 

I only have this client. My other computer is an iMac that acts as Retrospect Server.

 

I sometimes have to reboot the client in order for the server to find it (Error -530 on the server otherwise).

 

I don't have any proxy servers.

I had one VPN listed in my Network preferences that I just deleted. It was last used over a year ago (to connect to my former employer's network).

 

I only have an Airport Extreme router that both Macs connect to via Ethernet.

 

Attached is the log. retroclient.log.3.log.zip

 

It contains lots of entries like this:

2016-06-08T20:17:06: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T20:17:06: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T20:17:11: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T20:17:11: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T20:17:16: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T20:17:16: NetCheckNewInterfaces: found new address 10.0.1.3:0

It ends with:

2016-06-08T21:41:20: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T21:41:20: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T21:41:25: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T21:41:25: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T21:41:25: NetCheckNewInterfaces: found new address 10.211.55.2:0
2016-06-08T21:41:25: netCheckNewInterfaces: found new address 10.211.55.2:0
2016-06-08T21:41:25: Starting name service using 10.211.55.2:0
2016-06-08T21:41:25: iplud: failed address = 0.0.0.0
2016-06-08T21:41:25: Multicast advertising can not be initialized. Port 497 already in use (probably by RetrospectEngine).
2016-06-08T22:01:56: bindToValidBootPort: Unable to bind to valid boot port

So it crashes sometimes when failing "Multicast advertising".

But that is logged hours earlier at the start of that log file, apparently without ill effects.

2016-06-08T19:36:16: Client version is 12.5.0.111
2016-06-08T19:36:16: NetStart: starting interface thread
2016-06-08T19:36:16: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:16: netCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:16: ConnStartListen: starting thread ConnStartListen for 127.0.0.1:0
2016-06-08T19:36:21: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:30: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:30: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T19:36:30: netCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T19:36:30: Starting name service using 10.0.1.3:0
2016-06-08T19:36:30: iplud: failed address = 0.0.0.0
2016-06-08T19:36:30: Multicast advertising can not be initialized. Port 497 already in use (probably by RetrospectEngine).
2016-06-08T19:36:30: Turning client initiated discovery on
2016-06-08T19:36:36: IPNSRegister(0): registered: "CSLT MBP","4d1cba7186eca633","0.0.0.0:0"
2016-06-08T19:36:36: ConnStartListen: starting thread ConnStartListen for 10.0.1.3:0
2016-06-08T19:36:37: bindToValidBootPort: gServerPID has been initialized to 184
2016-06-08T19:36:41: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:41: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T19:36:41: Current IP Address 10.0.1.3
2016-06-08T19:36:41: --- UpdateClientStateWithMonitorEntered: clientState: 10
2016-06-08T19:36:46: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:46: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T19:36:51: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:51: NetCheckNewInterfaces: found new address 10.0.1.3:0
2016-06-08T19:36:56: NetCheckNewInterfaces: found new address 127.0.0.1:0
2016-06-08T19:36:56: NetCheckNewInterfaces: found new address 10.0.1.3:0

OK, I forgot I had installed the Retrospect Server for testing many months ago.

Is it enough to turn off Retrospect Engine and don't allow it to start? Or should I uninstall Retrospect Server completely? 

Hmmm? It's actually version 9.0.2, so I guess I want to uninstall it anyway. 

I'll do that. Thanks for replying.

[Lennart_T]

The 9.0.2 uninstaller was "damaged".

I downloaded a new 9.0.2 install package and that uninstaller was "damaged", too.

 

I right-clicked on the uninstaller and was able to extract the actual script file ("main.scpt"), which ran just fine. Now a reboot, just for good measure.

[eppinizer]

Hi Lennart,

 

From this information I have a feeling that the issue was a conflict between the client and the 9x engine running at the same time. We noticed an increase in the amount of problems caused by running both engine and client on the same machine right around when we released version 12. It wouldn't be to surprising if having an even older version of the engine could occasionally cause unusual behavior such as this. 

 

The crash was clearly related to enumerating your NIC list, and there having engine+client installed will have an effect in this area. Keep an eye on the client moving forward and let us know if you continue to see the behavior.