Raspberry 4 400, process exited with code 127

Iggy
Iggy
Joined: 19 Jan 21
Posts: 12
Credit: 22290
RAC: 0
Topic 224536

Hi. 

My first several work units errored out so I suspended work until I can get the problem figured out. Please advise if possible. Thank you.
 

 

<core_client_version>7.14.2</core_client_version>
<![CDATA[
<message>
process exited with code 127 (0x7f, -129)</message>
<stderr_txt>
../../projects/einstein.phys.uwm.edu/einsteinbinary_BRP4_1.06_arm-unknown-linux-gnueabihf: relocation error: ../../projects/einstein.phys.uwm.edu/einsteinbinary_BRP4_1.06_arm-unknown-linux-gnueabihf: symbol h_errno version GLIBC_PRIVATE not defined in file libc.so.6 with link time reference

</stderr_txt>
]]>

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4704
Credit: 17550336397
RAC: 6434310

You just ran into the same

You just ran into the same problem I have now on my RPi 3.  We used to get the beta 1.47 NEON application that worked fine.

But they just instituted a beta Android application and that took away our beta 1.47 NEON application as I think the project only allows ONE beta application per sub project.

And the stock 1.06 application never worked on ARMv7 or ARMv8 hardware. 

It only ever worked on ARMv6 hardware.

All my BRP4 work errors out now with the wrong 1.06 application.

Were screwed now unless the developers can sort out the scheduler to send us the correct application again.

Look at my host and you can see it produces valid work with the old beta 1.47 NEON application and nothing but instant errors with the 1.06 application the scheduler forced it onto.

Host 12773832

 

solling2
solling2
Joined: 20 Nov 14
Posts: 219
Credit: 1563293294
RAC: 47192

Keith Myers schrieb: You

Keith Myers wrote:

You just ran into the same problem I have now on my RPi 3.  We used to get the beta 1.47 NEON application that worked fine.

But they just instituted a beta Android application and that took away our beta 1.47 NEON application as I think the project only allows ONE beta application per sub project.

...

Host 12773832

There's yet another app which is doing fine. 

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4704
Credit: 17550336397
RAC: 6434310

They fixed the scheduler and

They fixed the scheduler and I'm getting 1.47 NEON tasks again.

 

Iggy
Iggy
Joined: 19 Jan 21
Posts: 12
Credit: 22290
RAC: 0

I added boinc-app-eah-brp

I added boinc-app-eah-brp from the available repository and it is crunching now.

 

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4704
Credit: 17550336397
RAC: 6434310

I'm trying this anonymous app

I'm trying this anonymous app now.  But I am NOT getting the kind of times that you and Mikey are getting.

It appears to run in the same times as the 1.47 NEON app.  I was hoping for the same kind of improvements you sort are getting.  Not for me unfortunately.

 

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828219831
RAC: 201209

Keith Myers wrote:I'm

Keith Myers wrote:

I'm trying this anonymous app now.  But I am NOT getting the kind of times that you and Mikey are getting.

It appears to run in the same times as the 1.47 NEON app.  I was hoping for the same kind of improvements you sort are getting.  Not for me unfortunately. 

One difference could be:

Keith's pi:

ARM
ARMv7 Processor rev 4 (v7l) [Impl 0x41 Arch 7 Variant 0x0 Part 0xd03 Rev 4]
(4 processors)
--- Linux Raspbian
Raspbian GNU/Linux 10 (buster) [5.4.83-v7+|libc 2.28 (Debian GLIBC 2.28-10+rpi1)]

 

Iggy's and Mikey's pi's:

ARM
ARMv7 Processor rev 3 (v7l) [Impl 0x41 Arch 7 Variant 0x0 Part 0xd08 Rev 3]
(4 processors)
--- Linux Raspbian
Raspbian GNU/Linux 10 (buster) [5.4.83-v7l+|libc 2.28 (Debian GLIBC 2.28-10+rpi1)]

 

Hmm no difference in those pi's:

One of Keith's tasks:

35,426 35,439 62

Binary Radio Pulsar Search (Arecibo) v1.47 () arm-unknown-linux-gnueabihf

 

 One of Iggy's tasks:

14,702 14,591 62 Binary Radio Pulsar Search (Arecibo) Anonymous platform

So it's the tasks not the device. Maybe you are still not running the Anonymous tasks yet Keith 

If it were me I would take another SD card and just start over from scratch and reload the Rpi software and then reload Boinc and the anonymous Einstein app and see if that works. I know you are more Linux savvy than I am but for me it would take less than 30 minutes from start to crunching. If that worked I would just redo the fast SD card the same way.

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4704
Credit: 17550336397
RAC: 6434310

No, these are the anonymous

No, these are the anonymous task I'm currently running.

https://einsteinathome.org/host/12773832/tasks/1/0

These are the two that I have successfully finished but ended up in Inconclusive status

https://einsteinathome.org/task/1061449815

https://einsteinathome.org/task/1061449776

So about an hour faster at just under 9 hours.  But certainly not the 4.5 or 5.5 hours you other guys are getting with the application.

I also see tons of errors in the Event Log about can't link to the slideshow files.

Midget

785    Einstein@Home    1/21/2021 7:21:40 PM    [error] Can't create link file projects/einstein.phys.uwm.edu/slideshow_einsteinbinary_BRP4_00    
786    Einstein@Home    1/21/2021 7:21:41 PM    [error] Can't create link file projects/einstein.phys.uwm.edu/slideshow_einsteinbinary_BRP4G_00    
787    Einstein@Home    1/21/2021 7:21:41 PM    [error] Can't create link file projects/einstein.phys.uwm.edu/slideshow_einsteinbinary_BRP4_06    
788    Einstein@Home    1/21/2021 7:21:42 PM    [error] Can't create link file projects/einstein.phys.uwm.edu/slideshow_einsteinbinary_BRP4G_05    
 

Must be something to do with the application installer putting the app_info.xml and binary into its own folder and creating symlinks back to the stock project folder.

In Projects I have a bogus einsteinathome.org folder along with the normal einstein.phys.uwm.edu folder.

There is also a new boinc-app-eah.brp folder in /var/lib along with the normal boinc-client folder.

 

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4704
Credit: 17550336397
RAC: 6434310

Yes, it would not be that

Yes, it would not be that much work.  A helluva lot more work on the Nano with its customized anonymous application.  I DO NOT want to mess with that.  It was a miracle for me to get the app running in the first place and I don't think I could ever recreate the conditions that allow it work now.

I have Universe running on the RPi3B+ also.  I had issues with that project getting the correct application also.

I'll let the other tasks run overnight and maybe one of them will get validated with a wingman.

 

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828219831
RAC: 201209

Keith Myers wrote:Yes, it

Keith Myers wrote:

Yes, it would not be that much work.  A helluva lot more work on the Nano with its customized anonymous application.  I DO NOT want to mess with that.  It was a miracle for me to get the app running in the first place and I don't think I could ever recreate the conditions that allow it work now.

I have Universe running on the RPi3B+ also.  I had issues with that project getting the correct application also.

I'll let the other tasks run overnight and maybe one of them will get validated with a wingman.

Does the Nano also use an SD card? If so you might want to find a way to clone it to another card just in case it goes bad some day, my wife who uses them in her camera says they go bad on her and they aren't usable anymore.

Yes I now see you are running some anonymous tasks let's hope for the best!!!

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4704
Credit: 17550336397
RAC: 6434310

Unfortunately, absolutely no

Unfortunately, absolutely no improvement in crunching times.  Exactly the same ~36K seconds for the anonymous application.

So since there is no benefit, I will just revert to the box stock 1.47 NEON application.

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.