| ID |
Date |
Icon |
Author |
Author Email |
Category |
OS |
ELOG Version |
Subject |
|
69887
|
Thu Aug 7 11:04:39 2025 |
 | Stefan Ritt | stefan.ritt@psi.ch | Question | Linux | V3.1.5-30a | Re: once a week we are having elogd segault? | Probably some very strange URL form nexpose to trigger a potential buffer overflow. If I get the precise URL which crashes elogd, I can reproduce and fix it.
Otherwise my usual advice: Run elogd behind an Apache proxy and do the authentication there. This way nexpose does not get to elogd, it will stop at the Apache (without the proper credentials).
Steafn
| mathew goebel wrote: |
|
We have since discovered that the security team is scanning the box in question once a week when the service crashes, with nexpose.
So if you see something similar then you might want to explore that.
| mathew goebel wrote: |
|
Jul 17 20:36:21 elog kernel: elogd[179095]: segfault at 7ffda4d82000 ip 00007f97033a1406 sp 00007ffda4d58c38 error 6 in libc-2.28.so[7f9703374000+1cd000]
Elog version ELOG V3.1.5-30ada1df
Running on a Rehdat 8 enterprise server
compiled with a Makefile change :: change -Wno-unused-result to -Wno-unused-value
Wondering if anyone has been seeing this?
|
|
|
|
69890
|
Sat Aug 16 22:19:03 2025 |
| Harry Martin | harrymartin772@gmail.com | Other | Linux | 3.1.5 | [CLOSED] Re: 3-level conditional not working | Closed. This "problem" was due to the same lack of document eyesight as https://elog.psi.ch/elogs/Forum/69889. Sorry for all the noise. New eyeglasses are forthcoming...
| Harry Martin wrote: |
|
Tried a variant, this time for NEW entries (not EDIT), which I meant to do to begin with; sorry if that was confusing.
[cond_test]
attributes = region, manufacturer, model
options region = american{us}, japanese{jp}, european{eu}
show attributes = region
{us} options manufacturer = ford, chrysler, gm
{jp} options manufacturer = toyota{toy}, nissan{nis}, mitsubishi{mit}
{eu} options manufacturer = audi, mercedes, bmw, volvo, fiat
{us} show attributes = region, manufacturer
{jp} show attributes = region, manufacturer
{eu} show attributes = region, manufacturer
{toy} options model = corolla, prius, miata
{nis} options model = sentra, murano
{mit} options model = starion, outlander, mirage
{toy,nis,mit} show attributes = region, manufacturer, model
I'm not sure if my multi-character conditional names are allowed or not, but they do seem to work... somewhat. Note that I even tried a couple different ways of specifying which attributes to show (single versus multiple OR conditionals). The problem now is that all of the attributes are always shown, seeming to ignore the conditionals entirely.
| Harry Martin wrote: |
|
Not sure if this is a bug or not; if it is not, kindly reclassify the category.
Following is a small example of what I will dub a 3-level, conditional logbook:
[cond_test]
attributes = region, manufacturer, model
show attributes edit = region
options region = american{1}, japanese{2}, european{3}
{1} options manufacturer = ford, chrysler, gm
{2} options manufacturer = toyota{4}, nissan{5}, mitsubishi{6}
{3} options manufacturer = audi, mercedes, bmw, volvo, fiat
{1,2,3} show attributes edit = region, manufacturer
{4} options model = corolla, prius, miata
{5} options model = sentra, murano
{6} options model = starion, outlander, mirage
{4,5,6} show attributes edit = region, manufacturer, model
This example is hardly complete, but it is sufficient to demonstrate the issue I am running into. This works until I select from the 2nd level (manufacturer); it won't display the dropdown for the model (I think my show attribute edit is supposed to display that). Is there a restriction on how many levels of conditional I can have? If there is, it would be disappointing because I am trying to do something similar to this example.
I hope there is something very simple I can do to correct this, or perhaps a completely different approach that accomplishes the same thing. I'd like to stick with the multi-level selection process as I have done here.
I am running elog 3.1.5 on Devuan Chimaera (approx. Debian Bullseye, sans systemd) in a Virtualbox VM.
|
|
|
|
69891
|
Mon Aug 18 03:14:51 2025 |
| Harry Martin | harrymartin772@gmail.com | Bug report | Linux | 3.1.5 | Re: collapse does not seem to work as expected | I am not sure if either "Sort Attributes" or the Quick Filter are the main issue, though they could be. I have implemented a work-around for now, but it really would be nice if this worked as expected. Or perhaps someone can provide an example of a logbook with working "collapse" functionality. Thank you for this nifty program; everything else aside, this tool helps keep my work in order.
| Harry Martin wrote: |
|
I also notice that collapse does not work properly if I have anything selected in the quick filter.
| Harry Martin wrote: |
|
Figured it out: I had "Sort Attributes" set on the logbooks that were failing to collapse correctly. At least, I think that is the distinction.
I think I can live with "Sort Attributes" disabled; I don't recall why I chose to set the "Sort Attributes" in the first place. But I still think this is an issue.
| Harry Martin wrote: |
|
I should add that this does not happen in every logbook. I'm looking into what the distinction is that causes this in certain logbooks and not in others.
| Harry Martin wrote: |
|
When I click on "collapse," even several times, I still see all the replies in each thread rather than just one entry representing each entire thread.
I am running elogd 3.1.5 on devuan chimaera (bullseye) which I built from source. I had the same results with 3.1.4 and 3.1.3.
I tried to disable the quick filter just to see if that impacted the behavior, but it didn't.
If there are any other tests I can use to narrow the problem, please suggest them. Without a working collapse feature, I cannot get a quick birdseye view of, say, all open threads.
|
|
|
|
|
|
69893
|
Mon Sep 15 15:11:41 2025 |
| Stefan Ritt | stefan.ritt@psi.ch | Question | Linux | 3.1.5-1272bc14 | Re: [global] config still editable by admin of top group | You can have authentication via the Webserver or the ELOG internal one, but this is on a global level for all logbooks. You cannot mix this between logbooks. For that, you would have to run two instances of ELOG at two different ports.
Stefan
| Damian Goeldi wrote: |
|
The ETH physics department is running an ELOG behind an Apache reverse proxy:
ProxyPass / http://localhost:$port/ retry=0
ProxyPassReverse / http://localhost:$port/
ProxyAddHeaders off
Authentication is done on the Apache side using LDAP authentication, example:
<Location /demo>
Use PhysLDAP
AuthType Basic
AuthBasicProvider ldap
...
Require valid-user
RewriteEngine On
RewriteCond %{LA-U:REMOTE_USER} (.+)
RewriteRule . - [E=RU:%1,NS]
RequestHeader add X-Forwarded-User %{RU}e
</Location>
And all ELOGs use the following config:
[demo]
Authentication = Webserver
For the PSI-Praktikum we had to create a logbook that is accessible without an ETH-Account. A new logbook was added, which is not authenticated via the proxy, but the ELOG internal authentication. In order to grant access to the students, I was made admin for that logbook. The configuration is the following:
[PSI-Praktikum]
Authentication = File
Password file = /home/wwwelog/private/password/psi-praktikum.xml
Admin user = damian
In order to prevent my user from editing the global configuration, top groups according to https://elog.psi.ch/elog/config.html#groups were introduced, with one top group for all the proxy-authenticated logbooks, and a separate one for the Praktikum logbook. However, even after doing this, I am still able to edit the [global] section. Is there a way to prevent this? Or is it not possible to have a global section that is not accessible by the top group admins?
|
|
|
46
|
Mon Jul 1 13:56:12 2002 |
 | Mo | msheikh25@hotmail.com | | | | Options <attribute> = List???? | Hello,
Is it possible to have a drop down list with commas in the options.
For instance, if you want to have a attribute "CityState" you would
probably want a comma between the City and State name. Since the Options
are separated by commas, it turns into two different options. I dont know
if there is a way to do this or if this is a bug?
MO |
|
48
|
Wed Jul 3 18:36:28 2002 |
| Stefan Siegel | s_siegel@gmx.net | Question | | | possible to modify link in email notification | Hello,
I have a question concerning the following situation:
There is a one logbook with two possible acces methods (defined in config
file) one for readonly access and one for read/write access. Both entries in
the elogd.cfg file pointing to the same logbook.
When an allowed person enters a logbook entry with an email notification the
link to the message is given inside the email.
This link points to the read/write logbook entry.
Is it possible to alter the link in the email notification that it points to
the corresponding readlonly logbook entry?
Thank you for hints, best regards,
Stefan Siegel |
|
56
|
Mon Jul 8 17:18:05 2002 |
| Mo | msheikh25@hotmail.com | Bug report | | | problem saving elogd.cfg | In version 2.0.4 of eLog I was having some problems with saving my
configuration file once I edited it on the web using the config. command.
I keep getting the message "Cannot open file %s: elogd.cfg". I made sure
the file was in the right directory. I dont know if I am doing something
wrong or if its a bug?
Mo. |
|
58
|
Mon Jul 8 21:01:00 2002 |
| Christian Korschan | Christian.Korschan@CTBTO.ORG | Bug report | | | Compiling elog 2.0.4 on Solaris | Comiling elogd 2.0.4 under Solaris obviously fails with
gcc -o elog elog.c -lsocket -lnsl
gcc -o elogd elogd.c -lsocket -lnsl
elogd.c: In function `ss_file_find':
elogd.c:1372: `alphasort' undeclared (first use in this function)
elogd.c:1372: (Each undeclared identifier is reported only once
elogd.c:1372: for each function it appears in.)
make: *** [elogd] Error 1
because alphasort is an ucb'ism which does not exist under SYSV
... not very portable :-( |
|