Sounds right. Tell us how it went.

. . Sadly ... not well.

. . When I tried to install the video drivers in recovery mode the install file threw up an error about not being able to use /tmp (of course, because it didn't exist). But not a problem I did it the long way through "console".

. . When I was editing app_info I used TBars posted app_info as a guide .... DAAAAHHHHHZZZZZ! I forgot he named his app differently to yours, another ghosted cache, you are not alone Keith :) But a simple enough task to correct the file name entries. So now it loaded some work and started to run ... but ... all compute errors. I quickly suspended BOINC and shut it down. I restored the original app_info and now working again as CUDA80. But I have a cache full of ghosted tasks to recover. And no explanation why it didn't work.

Stephen

:(

I think you need to remove the old CUDA80 library files from app_info.. And get the app_name correct. I had no troubles with converting my farm to the new 0.97 special app on second try.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I think you need to remove the old CUDA80 library files from app_info.. And get the app_name correct. I had no troubles with converting my farm to the new 0.97 special app on second try.

. . Hi Keith,

. . As I said, I used TBar's file as a guide and he had them removed so that is already done. My only mistake was in cutting and pasting his filename, since I downloaded Petri's link they didn't match and zzzork! No more cache. A simple retype to match Petri's filename and it loaded, recognised the app and downloaded new work, but it all fails to run, instant compute errors ...

. . At that point it is beyond my pay grade :( Something about cannot find some resource or other. But I was under the impression all needed resources were either in the app or in the video drivers.

Stephen

:(

http://setiathome.berkeley.edu/results.php?hostid=8222433&offset=0&show_names=0&state=6&appid=

What you failed to notice is that Keith and myself used PETRIS OTHER APP, the one for MAXWELLs.
You chose to use the OTHER PETRI APP, and didn't bother to look at the name difference.
IF, you would have had BOTH of Petri's Apps in the Folder, it wouldn't have GHOSTED the tasks. BOINC MUST FIND THE NAMED FILE.
If you have a named file in you app_info, it had Better be in the folder.
Your other problem Appears to be you didn't Enable Networking in Recovery Mode as per instructions.
If you don't Enable Networking, Nothing will be written to the disk.

Is this the same system of yours that's been borked for some time? Something about you can't update the Kernel?
I remember some problems you were having. Petri's App must have at least the same Kernel that was used to Compile the App.
Your Kernel may be too old, hence the Compute Errors. His Apps will Not work in Ubuntu 14.04.1, the one I use to compile with kernel 3.13. I had to change systems to use his App.
This Means your System is too Old, /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.20' not found
Looks as though it's time you finally fixed that system.

Well first requirement is a Nvidia driver with CUDA9.2 capability. So 396.45 or 396.51 is mandatory. The tasks you linked show a missing dependency on your system. Since I believe you are either running off a USB stick or minimal installation, you must not have the normal packages installed. You should open a Terminal. Type ldd and a space. With File Manager opened on the directory you have the 0.97 special app unpacked in, I assume either the /Downloads or /setiathome.berkeley.edu directory, click on the special app with the mouse and drag and drop it into the Terminal window. Click on the Terminal window again to make it have focus and hit Enter. The ldd command will print out all the dependencies that are needed by the 0.97 special app. The printed output at the top will show which packages are missing followed by all the packages dependencies that are satisfied. You need to install the packages that are missing. By the looks of your errored tasks, I would say you are missing the gcc6+ package libraries or gnu related.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

`GLIBCXX_3.4.21' is required. See stderr from your fauled tasks.

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

And I'm guessing (finding) that `GLIBCXX_3.4.21' can't/won't work with Ubuntu 14 ??

All you have to do is Update it to 14.04.5 and it will work.
The 14.04 system that isn't working hasn't been updated in some time.
Petri is running 4.2.0-42, you need at least that.

And I'm guessing (finding) that `GLIBCXX_3.4.21' can't/won't work with Ubuntu 14 ??

I'm sure it can be found or compiled. I must have it from somewhere.

This might be of help https://askubuntu.com/questions/575505/glibcxx-3-4-20-not-found-how-to-fix-this-error
It is not 3.4.2.1 but doing similar google searches might help.

I compile my programs with
gcc --version
gcc (Ubuntu 5.2.1-22ubuntu2) 5.2.1 20151010
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

I have been Asking.Ubuntu a lot here, LOL
Been through a bunch of those toolchain installs.

Description: Ubuntu 14.04.5 LTS
Release: 14.04

mine for ... strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep GLIBCXX
<snip> ....
GLIBCXX_3.4.18
GLIBCXX_3.4.19
GLIBCXX_3.4.20
GLIBCXX_3.4.21
GLIBCXX_3.4.22
GLIBCXX_3.4.23
GLIBCXX_3.4.24
GLIBCXX_3.4.25
GLIBCXX_DEBUG_MESSAGE_LENGTH

stderr ...
../../projects/setiathome.berkeley.edu/setiV0.97.linux_x64_10x0: relocation error: ../../projects/setiathome.berkeley.edu/setiV0.97.linux_x64_10x0: symbol _ZTVNSt7__cxx1119basic_istringstreamIcSt11char_traitsIcESaIcEEE, version GLIBCXX_3.4.21 not defined in file libstdc++.so.6 with link time reference

I'm going to look through that symbolic link again ...

Here is what I have...

root@Linux1:~/sah_v7_opt/Xbranch/client# ldd setiathome_x41zc_x86_64-pc-linux-gnu_cuda92     (<-- that is my latest executable, I have not updated the Makefile in a long time for the name)
	linux-vdso.so.1 =>  (0x00007ffc539e5000)
	libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f8aee5fd000)
	libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f8aee3f9000)
	librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f8aee1f0000)
	libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f8aede6e000)    <<<-------------- here is my libstdc++
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f8aedb66000)
	libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f8aed94e000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f8aed584000)
	/lib64/ld-linux-x86-64.so.2 (0x000055b0abbfd000)

root@Linux1:~/sah_v7_opt/Xbranch/client# gcc --version
gcc (Ubuntu 5.2.1-22ubuntu2) 5.2.1 20151010
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

root@Linux1:~/sah_v7_opt/Xbranch/client#  strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep GLIBCXX    <<------ and here is what is in it regarding GLIBCXX
GLIBCXX_3.4
GLIBCXX_3.4.1
GLIBCXX_3.4.2
GLIBCXX_3.4.3
GLIBCXX_3.4.4
GLIBCXX_3.4.5
GLIBCXX_3.4.6
GLIBCXX_3.4.7
GLIBCXX_3.4.8
GLIBCXX_3.4.9
GLIBCXX_3.4.10
GLIBCXX_3.4.11
GLIBCXX_3.4.12
GLIBCXX_3.4.13
GLIBCXX_3.4.14
GLIBCXX_3.4.15
GLIBCXX_3.4.16
GLIBCXX_3.4.17
GLIBCXX_3.4.18
GLIBCXX_3.4.19
GLIBCXX_3.4.20
GLIBCXX_3.4.21                                            <<<-------------- 3.4.21 is the last one I have
GLIBCXX_DEBUG_MESSAGE_LENGTH

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

So who is going to volunteer to develop a snap version of the 0.97 application installation and make a easy one click installation? That is what Richard keeps hinting at.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

....or, you could just wait around a few days until I've tested the version of V0.97, that I compiled in Ubuntu 14.04.1, long enough to feel safe about posting it.
It was compiled in an old enough version of 14.04 that it will work on almost ALL versions of 14.04, sorta like All the past versions I've posted.
It will probably even work on Stephen's system.
But, I just started testing it Yesterday...a few hours after receiving the source code.
It's the one running here, https://setiathome.berkeley.edu/results.php?hostid=6813106&offset=340
I'd like to see a few more days of testing.
It will also contain an actual app_info.xml...

Live booted Ubuntu doesn't have MidnightCommander. What happened with the world %)
EDIT: I observe very strange installation procedure.
Minimal install was chosen. So it installed... seems ALL. And now it REMOVES stuff. For example, it removed GNOME docs and now preparing to remove Majong game. Is it not common logic to NOT install everything on MINIMAL install? Even Windows at least pretends not to install smth that didn't selected for install...
EDIT2: Remmina (RDP client) asks to send usage statistics.. Is it really *nix OS or M$ won everywhere ???

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

What you failed to notice is that Keith and myself used PETRIS OTHER APP, the one for MAXWELLs.
You chose to use the OTHER PETRI APP, and didn't bother to look at the name difference.
IF, you would have had BOTH of Petri's Apps in the Folder, it wouldn't have GHOSTED the tasks. BOINC MUST FIND THE NAMED FILE.
If you have a named file in you app_info, it had Better be in the folder.

. . True but I had already said I knew the mistake I made there, I corrected that and I can recover the ghosts later.

Your other problem Appears to be you didn't Enable Networking in Recovery Mode as per instructions.
If you don't Enable Networking, Nothing will be written to the disk.

. . OK, that is something to note for the future. Never having used 'recovery mode' before I went back over the messages to get the details but I did not find the message with that detail in it. If I had the time and the patience I would go back over this thread and try to get all the important info like that together in a short series of messages. But like the saying says, we learn more from mistakes than from getting it right.

Is this the same system of yours that's been borked for some time? Something about you can't update the Kernel?
I remember some problems you were having.

. . Nope, that is the system with the 970's. I can update the kernel but any kernel later than release 96 and it won't boot from the flashdrive any more. So it stays at release 96.

Petri's App must have at least the same Kernel that was used to Compile the App.
Your Kernel may be too old, hence the Compute Errors. His Apps will Not work in Ubuntu 14.04.1, the one I use to compile with kernel 3.13. I had to change systems to use his App.
This Means your System is too Old, /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.20' not found
Looks as though it's time you finally fixed that system.

. . I actually asked Petri about kernel compatibility and he assured me he had compiled on a system running 14.04 so it should have worked. I guess there was more detail I should have know about but sadly I am no Linux guru. This system is running release 131 which is pretty much up to date. This system is the one on which I foolishly installed the repository version of BOINC. So upgrading to a later version of Ubuntu is a problem, I don't want to trash the host ID and create a new one. Catch 22's all over the place. If I choose the option in the update process to step up to a later Linux kernel, say 18.04 will it do a 'friendly' update without trashing everything/anything?

Stephen

<shrug>

Yes, that Minimal install sure is bass-ackwards. Install everything first . . . Then delete most of what you installed. Huh!? Canonical must have stock in storage companies to help exercise hard drives. I was disappointed in the Remmina data usage too. There are alternatives though.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Well first requirement is a Nvidia driver with CUDA9.2 capability. So 396.45 or 396.51 is mandatory. The tasks you linked show a missing dependency on your system. Since I believe you are either running off a USB stick or minimal installation, you must not have the normal packages installed. You should open a Terminal. Type ldd and a space. With File Manager opened on the directory you have the 0.97 special app unpacked in, I assume either the /Downloads or /setiathome.berkeley.edu directory, click on the special app with the mouse and drag and drop it into the Terminal window. Click on the Terminal window again to make it have focus and hit Enter. The ldd command will print out all the dependencies that are needed by the 0.97 special app. The printed output at the top will show which packages are missing followed by all the packages dependencies that are satisfied. You need to install the packages that are missing. By the looks of your errored tasks, I would say you are missing the gcc6+ package libraries or gnu related.

. . Yep you're right there are 2 missing libraries. I will hit that with Synaptic and see how that goes.

Stephen

<my bad>

And I'm guessing (finding) that `GLIBCXX_3.4.21' can't/won't work with Ubuntu 14 ??

. . Give the man the cigar ... All searches in Synaptic find nothing ...

Stephen

:(

All you have to do is Update it to 14.04.5 and it will work.
The 14.04 system that isn't working hasn't been updated in some time.
Petri is running 4.2.0-42, you need at least that.

. . This is not the machine that you think it is. This one gets updated regularly and works AOK (which is strange because it is quite an old C2D while the machine with the updated kernel issues is an i5-6600, the second newest machine I have, only 12 months old).

. . How do I tell exactly what version/revision of the kernel is running? In 'about Ubuntu' it only says 14.04 LTS?

Stephen

?

What was the output of the ldd Terminal run?

uname -r

prints out the kernel revision number.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)