Web-App doesn't start after updating to 2018-R1.1

[Alexandre_Cunha]

@389386:@Greg O'Lone

Out of curiosity, have you tried just running the binary from the command line to see if any other errors are generated?

Could you please tell me exactly how to do that?

[Greg_O_Lone]

Using a terminal, log in to the server, go to the directory where your app is and type

./devmsidatastore

It doesn’t get you anything except another check to see if something is missing.

[Alexandre_Cunha]

[quote=389413:@Greg O’Lone]Using a terminal, log in to the server, go to the directory where your app is and type

./devmsidatastore

It doesn’t get you anything except another check to see if something is missing.[/quote]

Segmentation fault

[AlbertoD]

[quote=389390:@Alexandre Cunha]my virtual machine at my mac runs ubuntu 14 server. both versions 2017 and 2018 runs fine.
…
the big problem here, on my point of view, is that i’m doing the same thing about 3 years, on same project. then only on this particular change, from 2017r3 to 2018r1.1, it just does not work! at all[/quote]

I know is frustrating and from your information, my guess is: there is something missing on CentOS 7.2.

As you know there are some changes in 2018r1.1 needed to run. Your virtual machine running ubuntu 14 has everything that 2018 needs and that’s why it runs fine.

I guess CentOS 7.2 is fine because Greg haven’t said anything about that version. Sorry I can’t help you more.

Edit: Just to try something else, I would install a virtual machine with CentOS 7.4 or 7.5, and test there, maybe there is something else needed from the server side and not just the libunwind library?

[Alexandre_Cunha]

@389450:@Alberto De Poo

Just to try something else, I would install a virtual machine with CentOS 7.4 or 7.5, and test there, maybe there is something else needed from the server side and not just the libunwind library?

I’m doing this about a week or more. Didn’t discover anything!

Does not work!!!

I’m very frustrated.

[Greg_O_Lone]

There’s nothing wrong with using 2018r1.1 with CentOS 7.2 as far as I know. That segmentation fault is definitely the problem.

I suggest that you open a second terminal and run

tail -f /var/log/messages

Then try running it from the command line again.

@Alexandre Cunha — Before we go beyond that, what’s your experience level with managing Linux servers? In other words, is this something you do every day?

[Alexandre_Cunha]

i’l[quote=389454:@Greg O’Lone]There’s nothing wrong with using 2018r1.1 with CentOS 7.2 as far as I know. That segmentation fault is definitely the problem.

I suggest that you open a second terminal and run

tail -f /var/log/messages

Then try running it from the command line again.

@Alexandre Cunha — Before we go beyond that, what’s your experience level with managing Linux servers? In other words, is this something you do every day?[/quote]

I’ll ask to support service to do that, because it’s hosted on a VPS.

Well, i can’t say i’m 100% linux expert, but i work with it (unix) since first Mac OS 10, on 2001…
(maybe i don’t know some linux tricks, but just because i never needed to learn)

But I do a lot of things on server, including shell scripts, installations, updates, etc…

I’m a certified mysql dba. i’m also a programmer since 1979…
(Fortran, Cobol, Assembler, Microsoft Basic, .Net, ASP, java, javaScript, php, etc…)

So i can say to you, i’m not 100% idiot. Maybe 10%…

And if i’m here trying to fix / understand this, is because i tried everything i could.

But Greg, don’t tell me the problem is on me!
I checked the documentation of 2018 release to make sure i didn’t forget something.
It’s all installed accordly.

Really want to fix this to bury the xojo’s 2017 version.

[Alexandre_Cunha]

[i]Hello,

In terminal 1 executing devmsidatastore* resulted in a segmentation fault.

root@host [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ls
./ OFF/ devmsidatastore* devmsidatastore\ Resources/ emails/ scripts/
…/ config.cfg devmsidatastore\ Libs/ devmsidatastore.cgi* images/

root@host [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ./devmsidatastore
Segmentation fault

root@host [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ./devmsidatastore
Segmentation fault

Terminal 2: tailing /var/log/messages. The ftp logs shown below were just a coincidence. Nothing was logged in /var/log/messages when I tried to execute ./devmsidatastore

root@host [~]# tail -f /var/log/messages

May 27 16:59:04 host pure-ftpd: (?@127.0.0.1) [INFO] New connection from 127.0.0.1
May 27 16:59:04 host pure-ftpd: (?@127.0.0.1) [INFO] __cpanel__service__auth__ftpd__X8XszqPhyYHJYWVm_gVewG2ScJlqN437fshomfnaDrFy35S6VvnQ04CnX20BD6co is now logged in
May 27 16:59:04 host pure-ftpd: (__cpanel__service__auth__ftpd__X8XszqPhyYHJYWVm_gVewG2ScJlqN437fshomfnaDrFy35S6VvnQ04CnX20BD6co@127.0.0.1) [INFO] Logout.[/i]

[RenéLandscheidt]

Hello @Alexandre Cunha,

is your App using Plugins?

You could also use ldd to check the needed libs for the libs (inside the Lib Folder of your App).
If you are using Plugins there is a chance that the Plugin-Libs need some not installed librarys.

For Example i have Problems on an Linux System with one of the eCrypt-Plugins from Einhugur. Especially the eCryptCompression Plugin if i try to use it on my OpenSUSE Mate 42.3 with Xojo even Xojo itself Segfaults (even with the latest eCrypt Plugin Version).

At the Moment i’m setting an VM with openSUSE 42.3 up to test if that is a Problem specially with my Installation or if it is something more - then i will be reporting this as a Bug to Einhugur.

[Greg_O_Lone]

[quote=389522:@RenLandscheidt]You could laso use ldd to check the needed libs for the libs (inside the Lib Folder of your App).
If you are using Plugins there is a chance that the Plugin-Libs need some not installed librarys.[/quote]
Scroll up and youll see that hes already done that.

[Greg_O_Lone]

[quote=389519:@Alexandre Cunha][i]Hello,

In terminal 1 executing devmsidatastore* resulted in a segmentation fault.

root@host [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ls
./ OFF/ devmsidatastore* devmsidatastore\ Resources/ emails/ scripts/
…/ config.cfg devmsidatastore\ Libs/ devmsidatastore.cgi* images/

root@host [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ./devmsidatastore
Segmentation fault

root@host [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ./devmsidatastore
Segmentation fault

Terminal 2: tailing /var/log/messages. The ftp logs shown below were just a coincidence. Nothing was logged in /var/log/messages when I tried to execute ./devmsidatastore

root@host [~]# tail -f /var/log/messages

May 27 16:59:04 host pure-ftpd: (?@127.0.0.1) [INFO] New connection from 127.0.0.1
May 27 16:59:04 host pure-ftpd: (?@127.0.0.1) [INFO] __cpanel__service__auth__ftpd__X8XszqPhyYHJYWVm_gVewG2ScJlqN437fshomfnaDrFy35S6VvnQ04CnX20BD6co is now logged in
May 27 16:59:04 host pure-ftpd: (__cpanel__service__auth__ftpd__X8XszqPhyYHJYWVm_gVewG2ScJlqN437fshomfnaDrFy35S6VvnQ04CnX20BD6co@127.0.0.1) [INFO] Logout.[/i][/quote]
You missed my point. You need to have two terminals logged into the server at the same time. One that’s running the tail command and the other that’s starting your binary. The reason for the segmented fault will often appear in the window where the tail command is running.

[RenéLandscheidt]

I see only the ldd for the App itself.

ldd devmsidatastor

not for Example the ldd for the Libs itself:

ldd /home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore/./devmsidatastore Libs/XojoConsoleFramework64.so

and so on. Especally if one uses Third Parity Plugins (like MBS for Example)

[Greg_O_Lone]

[quote=389518:@Alexandre Cunha]i’l

I’ll ask to support service to do that, because it’s hosted on a VPS.

Well, i can’t say i’m 100% linux expert, but i work with it (unix) since first Mac OS 10, on 2001…
(maybe i don’t know some linux tricks, but just because i never needed to learn)

But I do a lot of things on server, including shell scripts, installations, updates, etc…

I’m a certified mysql dba. i’m also a programmer since 1979…
(Fortran, Cobol, Assembler, Microsoft Basic, .Net, ASP, java, javaScript, php, etc…)

So i can say to you, i’m not 100% idiot. Maybe 10%…

And if i’m here trying to fix / understand this, is because i tried everything i could.

But Greg, don’t tell me the problem is on me!
I checked the documentation of 2018 release to make sure i didn’t forget something.
It’s all installed accordly.

Really want to fix this to bury the xojo’s 2017 version.[/quote]
I wasn’t trying to suggest anything. I like to ask about experience because it helps me figure out how detailed I need to be in my explanations.

[brian_franco]

alexandre

have you tried creating a new web project with 2018 R1.1 and see if that will run?

[Alexandre_Cunha]

@389526:@Greg O'Lone

I wasn’t trying to suggest anything. I like to ask about experience because it helps me figure out how detailed I need to be in my explanations.

Yes, i know.

As we are from different countries, different cultures and different languages, i’ll appreciate if all you guys could by the more detailed as possible.

[Alexandre_Cunha]

[quote=389527:@brian franco]alexandre

have you tried creating a new web project with 2018 R1.1 and see if that will run?[/quote]

Yes, sure! Same problem!

(created a centos 7.2 vmware on my mac, and i’m having problems too. i’m redoing all the installation and tests to make sure )

[Alexandre_Cunha]

@389524:@Greg O'Lone

You missed my point. You need to have two terminals logged into the server at the same time. One that’s running the tail command and the other that’s starting your binary. The reason for the segmented fault will often appear in the window where the tail command is running.

Yes, i understood. I asked them to do from 2 terminals. I asked again. i’m waiting.

[Alexandre_Cunha]

Hi @Greg O’Lone

Here are my conversation with support team, from older to recent…

##### (support)

cP/ssd14-tx/16598 root@67.222.29.134 [/home/msidatastore/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ./devmsidatastore
Segmentation fault

Nothing is output to /var/log/messages

It does not run under the local user either.

–
msidatastore@msidatastore.com.br [~/dev.msidatastore.com.br/cgi-bin/devmsidatastore]# ./devmsidatastore
Segmentation fault

##### (me)

Just to be sure…

Did you run first “tail” on one terminal
And then on other, the ./devmsidatastore?

Both terminals opened?

##### (support)

Yes.

Two terminals, one with tail -f /var/log/messages to capture all recent output and/or occurring output to /var/log/messages.

Another terminal to execute the file directly via root and the cPanel user.

No messages were output do /var/log/messages when executed.

I would not expect these files to dump to /var/log/messages on cPanel servers.

##### (me)

got it.

where would these logs be generated?

can you do new tests to catch the error in the most appropriate place?

##### (support)

Errors would be typically dumped to local files within the same directory if the script was setup to do so.

However, no files are being created indicating that the file/script is not outputting its errors because it never reaches that point.

I attempted to check other logs, such as /usr/local/apache/logs/error_log – but there was no information in regards to the specified script.

I attempted to perform an strace on the process to determine if any information relevant came out of it, but there was nothing that stuck out to me.

–
msync(0x114e000, 4096, MS_ASYNC) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], ~[KILL STOP RTMIN RT_1], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[KILL STOP RTMIN RT_1], NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], ~[KILL STOP RTMIN RT_1], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[KILL STOP RTMIN RT_1], NULL, 8) = 0
msync(0x7fe5ff81a000, 4096, MS_ASYNC) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], ~[KILL STOP RTMIN RT_1], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[KILL STOP RTMIN RT_1], NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], ~[KILL STOP RTMIN RT_1], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[KILL STOP RTMIN RT_1], NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[RTMIN RT_1], ~[KILL STOP RTMIN RT_1], 8) = 0
rt_sigprocmask(SIG_SETMASK, ~[KILL STOP RTMIN RT_1], NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
msync(0x8395d52a4774000, 4096, MS_ASYNC) = 0
— SIGSEGV {si_signo=SIGSEGV, si_code=SI_KERNEL, si_addr=0} —
+++ killed by SIGSEGV +++

[Alexandre_Cunha]

It’s crazy, isn’t?

[AlbertoD]

Alexandre, do you have a sample project that work with 2017r3 and not with 2018r1.1 on CentOS 7.2?

Maybe someone could test a VM with CentOS 7.2 and your sample project and find more information?