Demo Discussion
Forum Config Examples Contributions Vulnerabilities
  Discussion forum about ELOG, Page 97 of 806  Not logged in ELOG logo
 Find |  Login |  Help 
New entries since:Thu Jan 1 01:00:00 1970
 Full | Summary | Threaded | Hide attachments  Show all entries    6444 Entries 
Goto page Previous  1, 2, 3 ... 96, 97, 98 ... 804, 805, 806   Next  
IDdown Date Icon Author Author Email Category OS ELOG Version Subject
  69067   Mon Nov 25 16:32:07 2019 Reply Stefan Rittstefan.ritt@psi.chRequestWindows3.1.3-fd7f1e2Re: Executing a shell command using elogd Windows service

Wow, having these color signal lamps really looke like a cute solution, made me laugh.

No, there is no other way than the Execute new thing. I have given up long time ago to use Windows services, because they are very hard to debug and very limited. So at our site everything runs under Linux.

Have you tried to specify the explicit path of your log file? Like Execute new = echo $Status > C:\Path\Last_status.log

Best,
Stefan

Frank Baptista wrote:

Sorry -- I somehow selected the wrong OS in my original message. Asleep at the wheel again.

Frank Baptista wrote:

Greetings!

We've been successfully running nearly a dozen separate logbooks on independent laptops -- all of them are running elogd as a Windows service. This works well, since I've also set up auto recovery options in the event that the service inadvertently stops.

Now, I have a need to place the value of an attribute of the latest log entry into a basic text file.  Of course, this works just fine if I have launched elogd -x as a normal executable, using Execute new = echo $Status > Last_status.log in my CFG file.  However, I would like to be able to do this using the Windows service which is running in the background. 

Is there another way to write the value of an attribute into a separate file?  If not, do I have to have a special build of ELOG in order to be able to enable the Windows service to execute shell commands?  For the record, these logbooks are running on secure laptops that are isolated onto their own network, and the user is unable to edit the CFG file.

In case you're wondering about the reason for the separate text file -- I've written a separate program which illuminates one of 4 different color signal lamps (mounted on a test station), based on the latest "Status" of the test station. (Running, Idle, Broken, Other).

I appreciate any guidance here -- this is a "big deal" here, as one glance over the floor gives us an idea of what's running (or not).

Thanks!
Frank

 

 

  69066   Mon Nov 25 16:25:21 2019 Reply Stefan Rittstefan.ritt@psi.chBug reportMac OSX3.1.4Re: Trouble on Catalina

The Makefile is kind of obsolete, I switched to CMake. So the Makefiel is mostly there as backup. To compile elog, do

elog$ mkdir build; cd build
elog/build$ cmake ..
elod/build$ make install

that should also find your openssl library, since CMake is more clever.

I personally have not yet switched to MacOS Catalina, since I have too many 32-bit programs running there and wait until all of these get updated. Only then I will be able to test elog under Catalina.

Stefan

John S. Haggerty wrote:

I decided to rebuild elog 3.1.4 in Catalina (MacOS 10.15.1), XCode 11.2.1.  As in previous builds, I needed to add to the Makefile pointers to openssl:

CFLAGS += -I/usr/local/opt/openssl/include

LIBS += -L/usr/local/opt/openssl/lib

The make builds cleanly, no errors, no warnings.  After make/make install, elogd segfaults immediately.  I saw the same behavior with the version in gitlab.  I kept my (very) old build and was able to make install it without recompiling and it still works.

I'll crack out the debugger when I have a chance to get more information, but has anyone else seen this problem?  

 

  69065   Sun Nov 24 21:10:28 2019 Reply Frank Baptistacaffeinejazz@gmail.comRequestWindows3.1.3-fd7f1e2Re: Executing a shell command using elogd Windows service

Sorry -- I somehow selected the wrong OS in my original message. Asleep at the wheel again.

Frank Baptista wrote:

Greetings!

We've been successfully running nearly a dozen separate logbooks on independent laptops -- all of them are running elogd as a Windows service. This works well, since I've also set up auto recovery options in the event that the service inadvertently stops.

Now, I have a need to place the value of an attribute of the latest log entry into a basic text file.  Of course, this works just fine if I have launched elogd -x as a normal executable, using Execute new = echo $Status > Last_status.log in my CFG file.  However, I would like to be able to do this using the Windows service which is running in the background. 

Is there another way to write the value of an attribute into a separate file?  If not, do I have to have a special build of ELOG in order to be able to enable the Windows service to execute shell commands?  For the record, these logbooks are running on secure laptops that are isolated onto their own network, and the user is unable to edit the CFG file.

In case you're wondering about the reason for the separate text file -- I've written a separate program which illuminates one of 4 different color signal lamps (mounted on a test station), based on the latest "Status" of the test station. (Running, Idle, Broken, Other).

I appreciate any guidance here -- this is a "big deal" here, as one glance over the floor gives us an idea of what's running (or not).

Thanks!
Frank

 

  69064   Sun Nov 24 20:29:24 2019 Question Frank Baptistacaffeinejazz@gmail.comRequestMac OSX3.1.3-fd7f1e2Executing a shell command using elogd Windows service

Greetings!

We've been successfully running nearly a dozen separate logbooks on independent laptops -- all of them are running elogd as a Windows service. This works well, since I've also set up auto recovery options in the event that the service inadvertently stops.

Now, I have a need to place the value of an attribute of the latest log entry into a basic text file.  Of course, this works just fine if I have launched elogd -x as a normal executable, using Execute new = echo $Status > Last_status.log in my CFG file.  However, I would like to be able to do this using the Windows service which is running in the background. 

Is there another way to write the value of an attribute into a separate file?  If not, do I have to have a special build of ELOG in order to be able to enable the Windows service to execute shell commands?  For the record, these logbooks are running on secure laptops that are isolated onto their own network, and the user is unable to edit the CFG file.

In case you're wondering about the reason for the separate text file -- I've written a separate program which illuminates one of 4 different color signal lamps (mounted on a test station), based on the latest "Status" of the test station. (Running, Idle, Broken, Other).

I appreciate any guidance here -- this is a "big deal" here, as one glance over the floor gives us an idea of what's running (or not).

Thanks!
Frank

Attachment 1: Signal_tower.jpg
Signal_tower.jpg
  69063   Fri Nov 22 02:55:50 2019 Warning John S. Haggertyhaggerty@bnl.govBug reportMac OSX3.1.4Trouble on Catalina

I decided to rebuild elog 3.1.4 in Catalina (MacOS 10.15.1), XCode 11.2.1.  As in previous builds, I needed to add to the Makefile pointers to openssl:

CFLAGS += -I/usr/local/opt/openssl/include

LIBS += -L/usr/local/opt/openssl/lib

The make builds cleanly, no errors, no warnings.  After make/make install, elogd segfaults immediately.  I saw the same behavior with the version in gitlab.  I kept my (very) old build and was able to make install it without recompiling and it still works.

I'll crack out the debugger when I have a chance to get more information, but has anyone else seen this problem?  

  69061   Thu Nov 21 18:10:28 2019 Reply David Walliswallis@aps.anl.govQuestionLinuxV3.1.4-ba84827Re: PAM authentication question

Hi Christoph,

Thanks for looking into this, if you can enable PAM + File, our users would be very happy!

The pam.d issue is probably related to CentOS/Red Hat, since our PAM expert warned me that it might be necessary.

Jan Christoph Terasa wrote:
David Wallis wrote:

I'm testing the PAM authentication feature, and have a couple questions, a suggestion, and a comment.

First the comment... it was pretty easy to get working, and is exactly what we need here, so thanks! Our PAM stack here is designed to allow logins with Active Directory, LDAP, or local accounts, so the PAM option preserves all of that.

The suggestion: In order to make it work, I had to add a symbolic link in /etc/pam.d:
    elogd -> system-auth
That might be considered for addition to the documentation (this was on Red Hat Enterprise Linux 7.7)

The questions:

  1. The docs indicate that "Self register" must be set to >= 1, but in the code (elogd.c, line 26453), if the PAM module is enabled, Self register is overriden to 0. The result is that no "register as new user" link is displayed on the login screen. Is that the intent?
  2. Related... can PAM and File authentication both be enabled? We have some logbooks that are used by both internal people (with an A/D account) and outside collaborators that get local elog accounts. This works with LDAP + File, can it work with PAM?

Thanks in advance!

 

David, thank you for reporting on your findings regarding the PAM feature. I will look into the points you mentioned:

0. On my machines (Debian testing and stable) I did not have to add anything to /etc/pam.d, but apparently Debian just uses implicit defaults then, and REHL might insist on using excplicit settings. Adding a hint in the documentation is certainly useful, thank your for the suggestion. Maybe elog should provide a pam.d config file (which can be installed/adapted by package maintainers for various OSes).

1.+2. If I remember correctly, I intentionally disabled registration when using the PAM backend, because users will register using their passwd/LDAP/NIS users, and new users can only be regustered using the appropriate tools for the authentication mechanism used. This might not be correctly reflected in the docs, I will check that. In the light of question 2., I can also re-investigate that policy, so that logins will check against both the elog user database and PAM. Self-registering can then be enabled again, and new registrees will go to the elog database. I will try to bringthe code in line with how LDAP works.

 

regards,

Christoph

 

  69060   Mon Nov 18 16:58:21 2019 Idea Roger Kaltroger.kalt@psi.chInfoLinux3.1.4Example scripts how to migrate or combine logbooks

Attached the shell scripts using awk and sed how I have migrated two separated logbooks into one single and how I re-adjusted certain attributes.

Attachment 1: run_modif.sh 
#!/bin/bash
# KR84, 28.10.2019

# the input files are the exported XML files from ELOG -> Finden


# search in XML for sring between <DATE> and </DATE>
# and replace with: <DATE> and </DATE><When> and </When>
echo "converting export_rf.xml ..."
cat ./export_rf.xml | 
    sed 's/<Personnel\(.*\)Personnel>/<Author\1Author>/g' |
    sed 's/<DATE\(.*\)DATE>/<DATE\1DATE>\n\t\t<When\1When>/g'   |
    sed 's/<Subject>\(.*\)<\/Subject>/<Title>\1<\/Title><Entry_Type><\/Entry_Type>/g' |
    sed '/<Machine>SwissFEL<\/Machine>/ {N;N; s/<Machine>SwissFEL<\/Machine>.*<Domain>OBLA<\/Domain>.*<Section>TRFCB/<Machine>OBLA<\/Machine>\n\t\t<Domain>All<\/Domain>\n\t\t<Section>TRFCB/g}'  |
    sed 's/<When>Mon, /<When>/g'                          |
    sed 's/<When>Tue, /<When>/g'                          |
    sed 's/<When>Wed, /<When>/g'                          |
    sed 's/<When>Thu, /<When>/g'                          |
    sed 's/<When>Fri, /<When>/g'                          |
    sed 's/<When>Sat, /<When>/g'                          |
    sed 's/<When>Sun, /<When>/g'                          |
    sed '/<When>.*<\/When>/{s/ Jan 20/.01./g}'                    |
    sed '/<When>.*<\/When>/{s/ Feb 20/.02./g}'                    |
    sed '/<When>.*<\/When>/{s/ Mar 20/.03./g}'                    |
    sed '/<When>.*<\/When>/{s/ Apr 20/.04./g}'                    |
    sed '/<When>.*<\/When>/{s/ May 20/.05./g}'                    |
    sed '/<When>.*<\/When>/{s/ Jun 20/.06./g}'                    |
    sed '/<When>.*<\/When>/{s/ Jul 20/.07./g}'                    |
    sed '/<When>.*<\/When>/{s/ Aug 20/.08./g}'                    |
    sed '/<When>.*<\/When>/{s/ Sep 20/.09./g}'                    |
    sed '/<When>.*<\/When>/{s/ Oct 20/.10./g}'                    |
    sed '/<When>.*<\/When>/{s/ Nov 20/.11./g}'                    |
    sed '/<When>.*<\/When>/{s/ Dec 20/.12./g}'                    |
    sed 's/ +0100<\/When>/<\/When>/g'                                     |
    sed 's/ +0200<\/When>/<\/When>/g'                                     > export_rf_modified.xml



#    sed 's/ Jan 20/.01./g'                              |
#    sed 's/ Feb 20/.02./g'                              |
#    sed 's/ Mar 20/.03./g'                              |
#    sed 's/ Apr 20/.04./g'                              |
#    sed 's/ May 20/.05./g'                              |
#    sed 's/ Jun 20/.06./g'                              |
#    sed 's/ Jul 20/.07./g'                              |
#    sed 's/ Aug 20/.08./g'                              |
#    sed 's/ Sep 20/.09./g'                              |
#    sed 's/ Oct 20/.10./g'                              |
#    sed 's/ Nov 20/.11./g'                              |
#    sed 's/ Dec 20/.12./g'                              |




# search in XML and add offset to all IDs because they shall not overlap when merged.
echo "converting export_llrf.xml ..."
cat ./export_llrf.xml           | sed 's/<Subject>\(.*\)<\/Subject>/<Entry_Type><\/Entry_Type>\n\t\t<Status><\/Status>\n\t\t<Title>\1<\/Title>\n\t\t<Inv_ID><\/Inv_ID>/g' > export_llrf_modified1.xml
cat ./export_llrf_modified1.xml | awk -F'\t\t<MID>|</MID>||' '{ if ($2!="") {print "\t\t<MID>"$2+2016"</MID>"} else { print $1} }' > export_llrf_modified2.xml
cat ./export_llrf_modified2.xml | awk -F'\t\t<REPLY_TO>|</REPLY_TO>||' '{ if ($2!="") {print "\t\t<REPLY_TO>"$2+2016"</REPLY_TO>"} else { print $1} }' > export_llrf_modified3.xml
cat ./export_llrf_modified3.xml | awk -F'\t\t<IN_REPLY_TO>|</IN_REPLY_TO>||' '{ if ($2!="") {print "\t\t<IN_REPLY_TO>"$2+2016"</IN_REPLY_TO>"} else { print $1} }' > export_llrf_modified.xml

rm -rf ./export_llrf_modified1.xml ./export_llrf_modified2.xml ./export_llrf_modified3.xml 

cat ./export_llrf_modified.xml | 
    sed '/<Machine>SwissFEL<\/Machine>/ {N;N; s/<Machine>SwissFEL<\/Machine>.*<Domain>Test Systems<\/Domain>.*<Section>TRFCB/<Machine>OBLA<\/Machine>\n\t\t<Domain>All<\/Domain>\n\t\t<Section>TRFCB/g}'  |
    sed 's/<When>Mon /<When>/g'                           |
    sed 's/<When>Tue /<When>/g'                           |
    sed 's/<When>Wed /<When>/g'                           |
    sed 's/<When>Thu /<When>/g'                           |
    sed 's/<When>Fri /<When>/g'                           |
    sed 's/<When>Sat /<When>/g'                           |
    sed 's/<When>Sun /<When>/g'                           |
    sed 's/-Jan-/.01./g'                              |
    sed 's/-Feb-/.02./g'                              |
    sed 's/-Mar-/.03./g'                              |
    sed 's/-Apr-/.04./g'                              |
    sed 's/-May-/.05./g'                              |
    sed 's/-Jun-/.06./g'                              |
    sed 's/-Jul-/.07./g'                              |
    sed 's/-Aug-/.08./g'                              |
    sed 's/-Sep-/.09./g'                              |
    sed 's/-Oct-/.10./g'                              |
    sed 's/-Nov-/.11./g'                              |
    sed 's/-Dec-/.12./g'                              |
    sed 's/ +0100<\/When>//g'                                     |
    sed 's/ +0200<\/When>//g'                                     |
    sed 's/<\/When>/:00<\/When>/g'                      |
    sed 's/<When>-:00<\/When>/<When><\/When>/g'                    > export_llrf_modified_datetime.xml

echo "export_llrf_modified_datetime.xml need manual edit for empty <When></When>"
Attachment 2: generate_import_llrf_fwd.sh 
#!/bin/bash
# KR84, 28.10.2019

# generate emtpy auto-fwd text for LLRF for 3100 entries and offset of 2000
echo "generated import_llrf_fwd.xml"
echo "<?xml version=\"1.0\" encoding=\"UTF-8\"?>" > import_llrf_fwd.xml
echo "<ELOG_LIST>" >> import_llrf_fwd.xml

declare -i ID
declare -i IDNEW

for ID in {1..3013}
do
    IDNEW=$ID+2016

    echo -e "\t<ENTRY>" >> import_llrf_fwd.xml
    echo -e "\t\t<MID>${ID}</MID>" >> import_llrf_fwd.xml
    echo -e "\t\t<DATE>Mon, 28 Oct 2019 20:00:00 +0200</DATE>" >> import_llrf_fwd.xml
#    echo -e "\t\t<DATE>28.10.2019 20:00:00</DATE>" >> import_llrf_fwd.xml
    echo -e "\t\t<ATTACHMENT></ATTACHMENT>" >> import_llrf_fwd.xml
    echo -e "\t\t<ENCODING>HTML</ENCODING>" >> import_llrf_fwd.xml
    echo -e "\t\t<When>28.10.2019 20:00:00</When>" >> import_llrf_fwd.xml
#    echo -e "\t\t<When>1572289200</When>" >> import_llrf_fwd.xml
    echo -e "\t\t<Author>Kalt Roger (KR84)</Author>" >> import_llrf_fwd.xml
    echo -e "\t\t<Machine>SwissFEL</Machine>" >> import_llrf_fwd.xml
    echo -e "\t\t<Domain></Domain>" >> import_llrf_fwd.xml
    echo -e "\t\t<Section></Section>" >> import_llrf_fwd.xml
    echo -e "\t\t<System></System>" >> import_llrf_fwd.xml
    echo -e "\t\t<Subsystem></Subsystem>" >> import_llrf_fwd.xml
    echo -e "\t\t<Subject>Automatic forward</Subject>" >> import_llrf_fwd.xml
    echo -e "\t\t<TEXT>&lt;meta http-equiv=&quot;refresh&quot; content=&quot;0; URL=&apos;https://elog-gfa.psi.ch/SwissFEL+RF/${IDNEW}&apos;&quot; /&gt;</TEXT>" >> import_llrf_fwd.xml
    echo -e "\t</ENTRY>" >> import_llrf_fwd.xml
done

echo "</ELOG_LIST>" >> import_llrf_fwd.xml
  69059   Sun Nov 17 14:55:11 2019 Reply Jan Christoph Terasaterasa@physik.uni-kiel.deQuestionLinuxV3.1.4-ba84827Re: PAM authentication question
David Wallis wrote:

I'm testing the PAM authentication feature, and have a couple questions, a suggestion, and a comment.

First the comment... it was pretty easy to get working, and is exactly what we need here, so thanks! Our PAM stack here is designed to allow logins with Active Directory, LDAP, or local accounts, so the PAM option preserves all of that.

The suggestion: In order to make it work, I had to add a symbolic link in /etc/pam.d:
    elogd -> system-auth
That might be considered for addition to the documentation (this was on Red Hat Enterprise Linux 7.7)

The questions:

  1. The docs indicate that "Self register" must be set to >= 1, but in the code (elogd.c, line 26453), if the PAM module is enabled, Self register is overriden to 0. The result is that no "register as new user" link is displayed on the login screen. Is that the intent?
  2. Related... can PAM and File authentication both be enabled? We have some logbooks that are used by both internal people (with an A/D account) and outside collaborators that get local elog accounts. This works with LDAP + File, can it work with PAM?

Thanks in advance!

 

David, thank you for reporting on your findings regarding the PAM feature. I will look into the points you mentioned:

0. On my machines (Debian testing and stable) I did not have to add anything to /etc/pam.d, but apparently Debian just uses implicit defaults then, and REHL might insist on using excplicit settings. Adding a hint in the documentation is certainly useful, thank your for the suggestion. Maybe elog should provide a pam.d config file (which can be installed/adapted by package maintainers for various OSes).

1.+2. If I remember correctly, I intentionally disabled registration when using the PAM backend, because users will register using their passwd/LDAP/NIS users, and new users can only be regustered using the appropriate tools for the authentication mechanism used. This might not be correctly reflected in the docs, I will check that. In the light of question 2., I can also re-investigate that policy, so that logins will check against both the elog user database and PAM. Self-registering can then be enabled again, and new registrees will go to the elog database. I will try to bringthe code in line with how LDAP works.

 

regards,

Christoph

Goto page Previous  1, 2, 3 ... 96, 97, 98 ... 804, 805, 806   Next  
ELOG V3.1.5-3fb85fa6