[ZBX-3219] User-defined scripts error if zabbix_server not running. Need a human description. Created: 2010 Nov 16 Updated: 2017 May 30 Resolved: 2012 Apr 11 |
|
Status: | Closed |
Project: | ZABBIX BUGS AND ISSUES |
Component/s: | Frontend (F) |
Affects Version/s: | None |
Fix Version/s: | 1.8.12, 1.9.8 (beta) |
Type: | Incident report | Priority: | Minor |
Reporter: | Oleksii Zagorskyi | Assignee: | Unassigned |
Resolution: | Fixed | Votes: | 2 |
Labels: | trivial, usability | ||
Remaining Estimate: | Not Specified | ||
Time Spent: | Not Specified | ||
Original Estimate: | Not Specified |
Attachments: | 1.need_a_human_description.png |
Description |
I propose to make a special human description. Something like: |
Comments |
Comment by richlv [ 2011 Dec 01 ] |
message should mention other possibilities like incorrect server ip/dns or security environment like selinux (very common causes) |
Comment by Vjacheslav Shipillo (Inactive) [ 2011 Dec 05 ] |
How you think this can be like so ? <zalex> I think this message for user should contain as much details as possible. |
Comment by Vjacheslav Shipillo (Inactive) [ 2011 Dec 05 ] |
As i understand this should be inserted in "SCRIPT ERROR" place?? <zalex> I thinks yes. |
Comment by Oleksii Zagorskyi [ 2011 Dec 05 ] |
I think the line with the error description could be included in the message. (a) (b) (c) (d) |
Comment by Oleksii Zagorskyi [ 2011 Dec 06 ] |
Dev branch rev 23787 tested. (a) (b) - variant with some available LAN IP (d) - with IP instead of DNS Issue will be reopened. <SlaVa> RESOLVED <zalex>
-> REOPENED <zalex> |
Comment by Oleksii Zagorskyi [ 2011 Dec 06 ] |
(1) |
Comment by Pavels Jelisejevs (Inactive) [ 2011 Dec 06 ] |
(4) It's a bad idea to add <pre> to CObject::show(). CObject is also used to render HTML tags. <pavels> CLOSED. |
Comment by Pavels Jelisejevs (Inactive) [ 2011 Dec 06 ] |
(5) Please, format the code according to our guidelines. <pavels> Please review my formatting in r23848. |
Comment by Oleksii Zagorskyi [ 2011 Dec 06 ] |
(6) all strings have to be reviewed by someone experienced && experienced in English <dimir> AKA Rich :-D <zalex> yes, because SlaVa took several strings from the description of my testing cases. They were NOT the strings for the GUI, they were only my thought <zalex> at least the line "Server in "localhost" and daemon not running." should be reviewed. Original version (my ) was "Server is Localhost and daemon not running". is <-> in <pavels> CLOSED. <pavels> For some reason I've thought that you've done it already. My fault. <pavels> Currently the PHP errors are replaced as follows: <richlv> rewrote & fixed some of the messages - suggested to use the following PHP: Connection refused PHP: No route to host PHP: Connection timed out PHP: Name or service not known <pavels> We should probably add a default message, that doesn't cover these specific cases. <richlv> in that case we should simply print out whatever error we got from php <zalex> Suggested by Rich messages reviewed, excellent! Rich, are you sure in the "PHP: Connection refused -> ZBX: 4. Firewall is blocking the connection" ? <richlv> firewalls can reject or drop packets, you will see timeout in one case and, connection refused in another. i'd omit tcp, as zabbix connections always are tcp - if we go that way, message will become very complex <SlaVa>RESOLVED r24160 with "Firewall is blocking TCP connection" variant ... <zalex> missing quotes for sever name %s in S_NOT_RUN. Would bu useful to see it as Rich suggested. "localhost" name hardcoded for S_TIME_OUT <pavels> Still missing the quotes in S_NOT_RUN. <pavels> CLOSED. |
Comment by richlv [ 2011 Dec 07 ] |
(7) please, add some summary on what messages we have currently settled on, as i found this issue a bit hard to follow <pavels> Considering that we've add some specific socket errors to the code, maybe we should also use more specific errors instead of S_SCRIPT_ERROR_BIG? E.g if the socket error is "connection timed out" it hardly because the server daemon is not running. <pavels> Done in (6). CLOSED. |
Comment by Pavels Jelisejevs (Inactive) [ 2011 Dec 13 ] |
TESTED. |
Comment by Alexander Vladishev [ 2011 Dec 19 ] |
(8) when executing a script with an unaccessible server daemon: <pavels> Slava, don't forget to leave a RESOLVED message when fixing an issue. Please review my commit r24090. <pavels> CLOSED. |
Comment by richlv [ 2011 Dec 22 ] |
work on this issue happens in svn://svn.zabbix.com/branches/dev/ZBX-3219 |
Comment by Pavels Jelisejevs (Inactive) [ 2011 Dec 23 ] |
(9) If fsockopen throws a different error, we need to display the PHP error as is, not the S_SCRIPT_ERROR_DESCRIPTION. <pavels> My mistake, everything was ok. CLOSED. |
Comment by Pavels Jelisejevs (Inactive) [ 2011 Dec 27 ] |
TESTED. Please, review r24271 before merging. |
Comment by Vjacheslav Shipillo (Inactive) [ 2011 Dec 28 ] |
fixed in 1.8 r 24351; trunk in r 24355 |
Comment by richlv [ 2011 Dec 28 ] |
(10) looks like unused strings have been added to english locale file for trunk <richlv> there still are unused strings both in 1.8 & trunk english locales - it should be verified whether they should be removed or actually used <Sasha> REOPENED List of unused variables: <Sasha> CLOSED |
Comment by richlv [ 2011 Dec 28 ] |
(11) we changed one string variable in 1.8 - 'S_SCRIPT_ERROR' <richlv> > svn log -vc 24385 ..F....... the mentioned revision does not seem to solve this problem <Slava>
|
Comment by richlv [ 2011 Dec 28 ] |
(12) in trunk, some strings in english locale file are defined twice <richlv> trunk still has duplicate string definitions
|
Comment by richlv [ 2011 Dec 28 ] |
(13)
|
Comment by richlv [ 2011 Dec 28 ] |
(14) trunk code still uses some 'S_' variables that were modified - it should not (for example S_SCRIPT_ERROR, maybe others) |
Comment by richlv [ 2011 Dec 29 ] |
(15) 'Connection to Zabbix server "%s" faileds. |
Comment by richlv [ 2011 Dec 29 ] |
erm, wait. the "resolved" messages above claim that it's resolved in 1.8 & trunk branches. looks like actual commits were to recreated development branches, which i did not spot right away. (10) - (15) above should be retested |
Comment by Oleksii Zagorskyi [ 2012 Feb 10 ] |
(16) <Vedmak> This not seem to be regression. Where did we have multiline error messages? This probably can be fixed but then we need newlines in translation strings and research is needed if it will work at all. CLOSED |
Comment by Oleksii Zagorskyi [ 2012 Apr 03 ] |
vedmak, I figured it out! When this development was active I several times tested the dev branch. Later, the line breaks have been removed during development in rev 23848, but I forgot/miss this point, sorry. Btw, Fix version is incorrect, should be 1.8.11 <- we should be careful in such cases (Rich has been changed it to 1.8.12 after my reopening. *AND* the dev branch currently has some changes not merged to stable branches ! Additionally do not forget to delete " REOPENED |
Comment by richlv [ 2012 Apr 04 ] |
as discussed with pootle developer, newlines should be supported by using \n - so we could use that in these messages as well. "the dev branch currently has some changes not merged to stable branches" <- that part i don't like at all, though... how did that happen ? |
Comment by Alexey Fukalov [ 2012 Apr 11 ] |
dev branch: svn://svn.zabbix.com/branches/dev/ZBX-3219 Should be tested if all gettext tools support escaped double quotes. <zalex> frontend successfully tested, I like it. I suppose all will be fine with the Pootle as well. |
Comment by Oleksii Zagorskyi [ 2012 Apr 12 ] |
Just a note: a |
Comment by Toms (Inactive) [ 2012 Apr 16 ] |
TESTED |
Comment by Alexey Fukalov [ 2012 Apr 17 ] |
svn://svn.zabbix.com/trunk 26879 |