Sudden problems with using EPPC for remote commands

Yes it seems like that. I’m surprised that there aren’t more complaints.

I have a similar set-up, except the wurlitzer is a PowerPC. The Mavericks side sees all the apps running there, so my problem is not a problem. Yet.
I went through the error logs and found nothing. Now the pain of re-installing, testing, re-upgrading…

Ouch!

And “Choose remote Application” ?

I’ve never used choose remote application because I’m checking always the remote application from the local host. The process element of System Events is available “ but not documented “ also from the Finder. And the Finder is the only scriptable application which is running permanently on every Mac.

Thanks StefanK, but that script still doesn’t work for me. After I got a couple of authentication issues ironed out, it still throws an error -600, application not running error.

I’ve encountered this in the past, and rebooting the target machine always fixed it. But not today. :frowning:

My setup is 6 web servers that use AppleScript to process the access logs (rolled nightly at midnight) and send the pertinent data to an application (Helix Server) on ‘machine #7.’ Every night launchd faithfully triggers the script on each machine. This all works quite well, with just a few occasional hiccups.

But yesterday I had a power outage when a UPS overloaded (don’t ask), and "machine #7’’ went down. It rebooted OK when power was restored, but today I noticed that all but one of the remote machines was unable to contact the target machine. The console error reads “CFNetwork SSLHandshake failed (-4)” and running the script manually returns the “Application isn’t running (-600)” error.

(It’s really odd to me that one of the machines is still able to establish contact. All the rest can not and they all return the same -600 error.)

Trying to debug this, I fired up Script Debugger and created a simple “choose remote application” script. When I run this, the target machine (running Mountain Lion Server, BTW) shows either no applications or a bit of garbage text.

So it seems that something is now corrupted on the target machine. I’ve tried rebooting a few times, I’ve tried a turn “Remote Apple events” off/reboot/turn “Remote Apple events” on sequence, I’ve tried switching the “Allow access for” switch to “All users,” and I’ve tried clearing out the allowed users list and rebuilding it. Nothing is fixing it.

Next up is to run Disk Utility and see if that fixes anything. If that doesn’t help, I’ll clear out some caches and maybe think about upgrading it to Mavericks or Yosemite.

Hello.

Sometimes it helps to clean out the system caches, when things errs mystically, It might not help in your cases, but at least it won’t hurt. You can use a tool such as Onyx, to clean them out, (you need to reboot right afterwards, so that the system can rebuild the caches.)

Onyx cleared caches on both machines and rebooted, no help.

Well, that was the only shot I had for you really, as eppc events aren’t really my cup of tea.

I apologize if those questions sound stupid, but have you looked at the Console.app, while you are running the scripts that tries to start up things, I’d especially look for firewall and or ipfw entries. But I’d rather go through every error thrown, also things that may seem unrelated, like Dyld errors. I’d also go through the settings in the appfirewall from the preferences pane, but I don’t know exactly what I’d look for there.

HTH anyways.

Just out of curiosity, are both machines language settings set to English? Is the application’s bundle name localized like messages.app? When using remote events I address remote application by the localized bundle name rather than the actual application name. I think Yosemite automatically corrects this, at least it did for me when I had Yosemite. But if your system language is set to English or the application’s bundle name is not localized you don’t have to look into this.

Hello.

I am wondering about one more thing, and I’m sorry if this is a stupid question, but what kind of answer do you get when you ping the remote address and port?

I discovered another breakdown today, with a different pair of machines doing a different function. Between the two (first one described above) I think I know what the problem is: Security Update 2015-002 (https://support.apple.com/en-us/HT204413) includes a fix for the “FREAK” exploit:

I’m convinced this is what broke it, because I’ve run some tests on various machines here, and have found that as long as the machines are 10.7.x or lower, there is no error. But if either of the machines is 10.8.5 or higher, it fails. I guess the eppc protocol uses “ephemeral RSA keys.”

The AppleScript error (-600: Application not running) isn’t very helpful, but the one captured in Console (-4: CFNetwork SSLHandshake failed) is what has me convinced that Security Update 2015-002 is the culprit.

I proved it (to myself at least) by testing on a friend’s machine (running 10.9.5) where the Security Update is not yet installed. Here’s how you can confirm it with a single machine:

tell application "Finder" of machine "eppc://127.0.0.1" to activate

Run that…
with Remote Apple Events turned off: error “Remote access not allowed”
with Remote Apple Events turned on and:
Security Update installed: error “Application not running”
Security Update not installed: Finder activates (no error)

That seems pretty conclusive to me. Anybody got a data point to refute it?

Now I’m off to file a bug report with Apple.

Nicely done, mstrange!! That seems pretty definitive to me.

Thanks!
I’ve reported my findings to http://radar.apple.com, http://www.openradar.me, http://www.macintouch.com, and http://www.macissues.com – let’s hope they fix this quickly.

Thanks mstrange. Learned again something new :slight_smile: . I always assumed it wasn’t encrypted but it was an open and insecure network protocol. Maybe I have missed the memo.

I got a reply from Apple (!) stating that “Engineering has determined that your bug report (20226451) is a duplicate of another issue (20219161)…” so we now know they are aware of it. Let’s hope there’s a fix made available quickly. (But thoroughly tested this time!)

Hello.

I believe there are other SSL implementations, you can use out there, to replace Apple’s current implementation with, but, they are probably patched the same way, anyways, if this is critical for running your business, then I’d go to Apples Opensource directory for Yosemite. I’d reall also make an account at The MacPorts Project – Home and check out anything regarding any implementations of openSSL that they distribute, there, at least, you may find people among the maintainers of any openSSL packages, that you can ask questions you may have.

Who knows, they may even provide an openSSL version that works for you! :slight_smile:

HTH

(And I have started to fancy the idea of my own Wurlitzer! :wink: )

Duplication message is a good thing. Priority of solving bugs works a bit different for Apple than other software companies. For Apple the amount of duplicates determines the level of priority of a bug. While normal bug report systems try to avoid duplicate reports, for Apple you should definitely create as much as duplicates as you like.

Looks like this was fixed in Security Update 2015 004.

FANTASTIC!!

Thanks so much for the update, DougAdams! And thanks to everyone who helped diagnose this and submit the bug report to Apple.

I can’t tell you how happy I am that all my scripts work again!

I’m currently running El Capitan GM and still getting the “Application not running” error, when the script is run from Script Editor and silently failing when run from Script Menu. In Yosemite the AEServer process would spike on the remote machine, now in El Capitan it no longer spikes.

Any help is greatly appreciated.