[rt-users] mod_fcgid: read data timeout leads to SIGKILL
New instance of RT 4.2.2 on CentOS 6.5 with Apache 2.2.15-29 and FCGI 2.3.9-1. Everything runs fine for the most part, however, intermittently the following log entries show up and I get an internal server error. [Fri Jan 31 22:48:36 2014] [warn] [client 192.168.236.64] mod_fcgid: read data timeout in 40 seconds, referer: http://rt.example.com/Ticket/Display.html?id=2 [Fri Jan 31 22:48:36 2014] [error] [client 192.168.236.64] Premature end of script headers: rt-server.fcgi, referer: http://rt.example.com/Ticket/Display.html?id=2 FastCGI: manager (pid 8954): received signal TERM FastCGI: manager (pid 8954): sending TERM to PIDs, 8960 [Fri Jan 31 22:48:46 2014] [warn] mod_fcgid: process 8954 graceful kill fail, sending SIGKILL FastCGI: server (pid 8960): safe exit after SIGTERM I've looked around and while there are various posts regarding the premature end of script headers I have yet to find one that mentions read data timeouts or SIGKILLs. Could the server simply be running slow? Load averages are sub-1. Restarting httpd doesn't show any errors. It's a virtual machine so I suppose the underlying hardware might be causing performance issues. -Mathew "When you do things right, people won't be sure you've done anything at all." - God; Futurama "We'll get along much better once you accept that you're wrong and neither am I." - Me
Re: [rt-users] RT 4.2 change scrip stage
On Fri, Jan 31, 2014 at 04:50:15PM +0100, Christian Loos wrote: > How can I delete a scrip? > The Method is there but I can't find anywhere a delete button for > scrips, even by checking the HTML source code. I have a vague recollection that the argument was made that deleting scrips was 'dangerous' because there's no way to get them back, so we were going to move to a Disabled/Enabled model like most objects in RT. Unfortunately, while that is a good argument, there are some UI failures supporting it. I'll file a bug about it. -kevin pgpiADzoHhsEf.pgp Description: PGP signature
Re: [rt-users] taking 55 plus seconds to return from responding to tickets
On Fri, Jan 31, 2014 at 04:27:47PM -0500, Lois Bennett wrote: Reordering - >By using the command line to create a ticket I have tracked it down to a > single call using the >Devel::NYTProfmodule. > >perl -d:NYTProf rt create -t ticket set subject="Testing" queue=19 > owner=salbb All you've done here is profile the REST interface. You're not *actually* profiling the RT codebase, this is why all your time is spent in Net::HTTP. >It has been running fine. Monday a colleague shutdown and cloned the RT > server in order to >give it a new name in the virtual center server. >The only change that was made to the server I believe was the ether net > address changed. I had >to start httpd and sendmail manually because they had not been set to auto > start. I did that >with the chkconfig command for sendmail and for httpd. Everything seemed > to come up fine >but. Your DNS is probably broken. When this happens, sendmail begins blocking when sending email. Step 1 - fix your DNS Step 2 - if you're running sendmail brand sendmail, read about and enable this configuration so that RT doesn't have to sit around and wait while sendmail mucks about sending email. http://bestpractical.com/docs/rt/latest/RT_Config.html#SendmailArguments (the second paragraph). If fixing DNS doesn't fix the issue, please post the logs you mentioned. If you want to profile RT itself, you need to either integrate NYTProf into your apache configuration or run it on /opt/rt4/sbin/standalone_httpd which will give you a guaranteed single threaded test site. -kevin pgp3T698lornx.pgp Description: PGP signature
[rt-users] taking 55 plus seconds to return from responding to tickets
I am really stuck on this. I suspect as always it will be something quite simple but I can find the solution. In December I upgraded from 2.8.9 to 4.2.2 on a new VMware Centos 6 server. It has been running fine. Monday a colleague shutdown and cloned the RT server in order to give it a new name in the virtual center server. The only change that was made to the server I believe was the ether net address changed. I had to start httpd and sendmail manually because they had not been set to auto start. I did that with the chkconfig command for sendmail and for httpd. Everything seemed to come up fine but. We are experiencing a problematic slow down. Creating tickets and responding to tickets and closing tickets takes over a minute every time. You can't use a different window for RT in the same browser but you can access if you use a different browser. The behavior is correct in the end for all the tickets seem to be created and sent properly. By using the command line to create a ticket I have tracked it down to a single call using the Devel::NYTProfmodule. perl -d:NYTProf rt create -t ticket set subject="Testing" queue=19 owner=salbb I find that it is spending 55.4 seconds on one call. # spent 55.4s within Net::HTTP::Methods::CORE:sselect which was called 4 times, avg 13.8s/call: # 4 times (55.4s+0s) by Net::HTTP::Methods::can_read at line 298, avg 13.8s/call sub Net::HTTP::Methods::CORE:sselect; # opcode I have it set to debug but I don't understand the logs enough to find a fix. Can anyone help? Lois Bennett, MSEE Senior System Administrator Channing Division of Network Medicine, Brigham & Women's Hospital *A Teaching Affiliate of Harvard Medical School and Harvard School of Public Health*
Re: [rt-users] Numeric custom field values
Sorry I forgot to add - I'm not sure if the versions will make any difference in regards to this question, but I'm running RT 4.2.2 on Ubuntu 13.10 with MySQL 5.5.34 and Sphinx 2.1.3 Thanks, Brent From: rt-users-boun...@lists.bestpractical.com [mailto:rt-users-boun...@lists.bestpractical.com] On Behalf Of Parish, Brent Sent: Friday, January 31, 2014 2:52 PM To: rt-users@lists.bestpractical.com Subject: [rt-users] Numeric custom field values Hi. I have some custom fields to track numeric priority independent of the built-in priority field. This is an example of a search string I use: Queue = 'helpdesk' AND 'CF.{Helpdesk Priority}' > 15 But this returns tickets where CF.{Helpdesk Priority} is set to 2 That leads me to believe that the operation being preformed is a string comparison (where 2 is indeed greater than 15), versus a numeric one. So my question is: is there a proper syntax to use in the search builder to treat those custom field values as numeric? Thanks! Brent
Re: [rt-users] RT-IR Clean install - non-existent subroutine RT::Queue::HasRight
Thanks, rolled back to 4.0.19 and all works!! -- View this message in context: http://requesttracker.8502.n7.nabble.com/RT-IR-Clean-install-non-existent-subroutine-RT-Queue-HasRight-tp56468p56472.html Sent from the Request Tracker - User mailing list archive at Nabble.com.
[rt-users] Numeric custom field values
Hi. I have some custom fields to track numeric priority independent of the built-in priority field. This is an example of a search string I use: Queue = 'helpdesk' AND 'CF.{Helpdesk Priority}' > 15 But this returns tickets where CF.{Helpdesk Priority} is set to 2 That leads me to believe that the operation being preformed is a string comparison (where 2 is indeed greater than 15), versus a numeric one. So my question is: is there a proper syntax to use in the search builder to treat those custom field values as numeric? Thanks! Brent
Re: [rt-users] 'negative' Business Hours in international environments
On 2014-01-29 16:33, Vitalik Yahoo wrote: >> I am using: Business-Hours-0.10_01 and RT-Extension-SLA-0.07. >> I have a follow-up question to this thread: http://www.gossamer-threads.com/lists/rt/users/106929 [1] >> I would like to implement this syntax but in our 24/7 operation Monday actually starts 7:00pm Sunday in order to align actually with the Shifts around the globe. First shift is APJ because the day begins there, i..e Monday morning begins in APJ at 7:00PM ET which is still Sunday in US. >> Is there a way to have 'Monday' Start on Sunday 7:00pm ET? i.e. 'Monday', Start => '-05:00',End=>'19:00'} >> %RT::ServiceBusinessHours = ( >> >> 'weekday' => { >> 0 => { Name => 'Sunday', Start => '19:00', End => '24:00' }, >> 1 => { Name => 'Monday', Start => '00:00', End => '24:00' }, >> 2 => { Name => 'Tuesday', Start => '00:00', End => '24:00' }, >> 3 => { Name => 'Wednesday', Start => '00:00', End => '24:00' }, >> 4 => { Name => 'Thursday', Start => '00:00', End => '24:00' }, >> 5 => { Name => 'Friday', Start => '00:00', End => '19:00' }, >> holidays => [qw(01-01 12-25 2014-05-26 2014-07-04 2014-09-01 2014-11-27 2014-11-28 )], >> }, >> ); Would it be possible to simply run your entire RT server in the APJ timezone, and work around the problem that way? -Adam athom...@athompso.net Links: -- [1] http://www.gossamer-threads.com/lists/rt/users/106929
Re: [rt-users] RT-IR Clean install - non-existent subroutine RT::Queue::HasRight
On Fri, Jan 31, 2014 at 10:45:55AM -0800, dajackman wrote: > Hello all. I am installing RT and RT-IR on a clean RHEL 6.5 machine. > Installed RT without any problems. Logged in to the web interface using the > default username password, no problems. Installed RT-IR 3.0.0. When I > restart apache and hit the site I get the following error: > rt-4.2.2 > RT-IR-3.0.0 RTIR 3.0 is for RT 4.0, not RT 4.2. There is no current RTIR release available that supports RT 4.2. If you're going to use RTIR, you may prefer the rtir mailing list, where this has been answered. -kevin pgpS4lrJXrGGd.pgp Description: PGP signature
[rt-users] RT-IR Clean install - non-existent subroutine RT::Queue::HasRight
Hello all. I am installing RT and RT-IR on a clean RHEL 6.5 machine. Installed RT without any problems. Logged in to the web interface using the default username password, no problems. Installed RT-IR 3.0.0. When I restart apache and hit the site I get the following error: [30936] [Fri Jan 31 18:29:47 2014] [critical]: Can't wrap non-existent subroutine RT::Queue::HasRight at /opt/rt4/sbin/../lib/RT.pm line 732. Compilation failed in require at /opt/rt4/sbin/../lib/RT.pm line 732. (/opt/rt4/sbin/../lib/RT.pm:393) Can't wrap non-existent subroutine RT::Queue::HasRight at /opt/rt4/sbin/../lib/RT.pm line 732. Compilation failed in require at /opt/rt4/sbin/../lib/RT.pm line 732. I removed everything from /opt/rt4, dropped the database and re-installed, same problem. No problems or warnings seen during the install. If I remove Set(@Plugins, 'RT::IR'); from RT_SiteConfig.pm, I can log back. Seems to be something with RT-IR. A few version numbers: Red Hat Enterprise Linux Workstation release 6.5 (Santiago) This is perl, v5.10.1 (*) built for x86_64-linux-thread-multi Apache/2.2.15 mysql-server-5.1.71-1 rt-4.2.2 RT-IR-3.0.0 Anyone have any ideas? Thanks in advance. -dajackman -- View this message in context: http://requesttracker.8502.n7.nabble.com/RT-IR-Clean-install-non-existent-subroutine-RT-Queue-HasRight-tp56468.html Sent from the Request Tracker - User mailing list archive at Nabble.com.
Re: [rt-users] How get Current user in scrip
On Fri, Jan 31, 2014 at 11:23:28AM +0400, Arkady Glazov wrote: >How i can get user id or name that login to web RT? In scrip > $self->CurrentUser show only >'RT_System' always (on any login). $self->TransactionObj->CreatorObj is the RT::User object of the person who made the change (transaction) which triggered the Scrip. -kevin pgpi6tI9MMmzW.pgp Description: PGP signature
Re: [rt-users] RT 4.2 change scrip stage
Another one: How can I delete a scrip? The Method is there but I can't find anywhere a delete button for scrips, even by checking the HTML source code. Chris
[rt-users] Creating Custom Dashboard - 'New Articles'
Hi im looking to create a new dashboard for users that will display all the new articles that have been created/updated in the last month or so. i have created dashboards for tickets before but never Articles so im struggling with the search terms required as the simple search seams geared toward tickets in 4.0.12 does anyone have any pointers? regards scott -- View this message in context: http://requesttracker.8502.n7.nabble.com/Creating-Custom-Dashboard-New-Articles-tp56465.html Sent from the Request Tracker - User mailing list archive at Nabble.com.