[10:00] * Now talking on #zenoss
[10:00] * Topic for #zenoss is: Zenoss Development will be here Thursday August 28 at 11am EST
[10:00] * Topic for #zenoss set by mrayzenoss at Thu Aug 14 16:06:33 2008
[10:01] Good morning.
[10:01] * jplouis (n=jplouis@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:01] hey matt
[10:02] jplouis and I are ready for questions, other Zenossians are lurking
[10:03] * _chris_ (n=chris@75-139-111-215.dhcp.mant.nc.charter.com) has joined #zenoss
[10:03] oh, I forgot the devs are here today..
[10:03] just one
[10:03] so far
[10:03] im guessing there was no progress on the whole RRD gaps/zenperfsnmp problems..
[10:03] so hopefully when I transition to enterprise this problem will go away..
[10:03] well, the same people will be working on them
[10:04] Chet and EAD were on that
[10:04] * bedwards (n=bedwards@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:04] * cluther (n=cluther@ip72-192-244-133.dc.dc.cox.net) has joined #zenoss
[10:04] speaking of the devil
[10:05] Hey Chet, jb was just asking about http://dev.zenoss.com/trac/ticket/3536
[10:05] wait.. I wasn't the devil, was I?
[10:06]
[10:08] * Erik-Zenoss (n=edahl@74.93.194.82) has joined #zenoss
[10:09] * Erik-Zenoss has quit (Client Quit)
[10:09] * Erik-Zenoss (n=edahl@74.93.194.82) has joined #zenoss
[10:10] * bwhichard (n=bwhichar@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:10] * bootay (n=bootay@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:10] * Erik-Zenoss has quit (Client Quit)
[10:11] any questions?
[10:11] seems pretty quiet today
[10:11] I have systems being shown in reports when using systems organizer in which they no longer exist in the organizer the particular report is using.
[10:12] mrayze: is there any way i can have the status of zencommand based checks be shown in the device status screen when they are not in error condition? we would like the reassuring green pill
[10:12] The device was removed from one organizer and placed in another, but it still is 'residually' reporting for the previous organizers
[10:12] this is 2.2.3
[10:12] teletype telexicon
[10:13] * dundel has quit (Read error: 113 (No route to host))
[10:15] I have been seeing strange errors in zenactions that seem to cause it to time out quite often when I have email rules set with counts.. Ver 2.2.3 x64
[10:15] kgoedtel: can you open a ticket with steps to reproduce?
[10:15] * paulh_ (n=paulh@74.93.194.82) has joined #zenoss
[10:16] I haven't been able to reproduce at will unfortunately
[10:16] i've viewed the organizer via zope manage to make sure i wasn't insane
[10:17] teletype: The individual device will show "green" on the device page, but you want it on the dashboard?
[10:17] teletype telexicon
[10:17] nope, on the device page status tab. zencommand based checks are only showing up in the status display if they are in an error state.
[10:17] teletype telexicon
[10:18] teletype: that sounds like a real bug, why don't you open a ticket
[10:18] kgoedtel: any steps and which reports would be helpful to track it down.
[10:18] ah, really? so i'm not crazy then. will do, thanks
[10:19] basically all i did was remove a device from one organizer under systems to another
[10:19] * cgibbons (n=cgibbons@75.148.215.110) has joined #zenoss
[10:19] i can provide screenshots
[10:19] which would probably illustrate the problem best
[10:19] teletype telexicon
[10:20] teletype: we'll verify the ticket. You're just running zencommand and nothing else, and it stays grey?
[10:20] Guyverix: didn't you have a forum thread for that?
[10:20] actually I opened a bug.
[10:21] was kinda curious if it was me, or a real bug
[10:21] i'm trying to use some custom nagios scripts via zencommand over ssh. the check works correctly, but if i look at the device status page, it does not show up at all. if i cause the script to return an error state, then the component is listed with the appropriate color pill under 'Other' category
[10:21] if i then cause the script to return an OK status, the component will no longer be listed
[10:23] Guyverix: ticket #?
[10:23] let me dbl check
[10:24] kgoedtel: screenshots and the report being run would be helpful
[10:24] lo matt el al.
[10:25] kgoedtel: This organizer/reporting behavior is expected. The organizers that an event is associated with are set at event time, not reporting time. So if an event is created for a device when it is in the /Mail system, then the device gets moved to the /Database system the event will remain a /Mail event.
[10:25] are there any examples of using SnmpPlugin.deviceProperties
[10:25] ?
[10:25] malbon: $ZENHOME/Products/DataCollector/plugins/zenoss/snmp/HRFileSystemMap.py
[10:26] bug 3593
[10:26] hrmf
[10:26] Guyverix: I just found it
[10:27] this report is a performance report
[10:27] Guyverix: and zenactions hangs and that's the last thing in the log?
[10:27] not an event report
[10:27] yes
[10:27] cluther: thanks.
[10:28] when I restart it, it says that it finished the action after usually over 1000 seconds
[10:28] Guyverix: I'll raise the priority, that doesn't sound like a medium bug
[10:28] kgoedtel: Which report(s)?
[10:28] ok, I pasted in what I think is the sql command being used, but I dont know if the h. is a place holder or not
[10:29] a custom one i created in multi-graph
[10:30] kgoedtel: ah.. ok.
[10:30] Guyverix: that ticket will get looked at, it looks like a good candidate to get fixed for 2.3
[10:31] i am getting some screen, have em in a bit
[10:31] cool!
[10:31] BTW, is there anything in the works to integrate Zenoss into external ticketing systems?
[10:32] Guyverix: Enterprise has a Remedy plugin
[10:32] Guyverix: I have done this using email and our ticketing system.
[10:32] There is an AlarmPoint ZenPack coming someday
[10:33] Ahh, OK.. So Remedy, via enterprise, and RT or the like via email..
[10:33] yeah, I'd like a good RT howto
[10:33] does the enterprise one for Remedy show the ticket that is associated with the device?
[10:34] I don't know the specifics, cluther?
[10:34] inside Zenoss itself, I mena
[10:34] Guyverix: Nope.. just Zenoss details in Remedy.
[10:35] hmm, ok
[10:35] kgoedtel: I can definitely reproduce that in the multi-graph reports.
[10:35] one thing about integrating them is that the zenpop doesn't have a mechanism to redirect to 'default' device for sorting out returns from the ticketing system of choice.
[10:36] is it possible to customize the link that comes with the email alert. It would work better for us if the url pointed to the IP of the host
[10:37] ${here\manageIp} ?
[10:37] oh great!
[10:38] cluther: I tried the getattr, however I cannot get a valid return for zDeviceTemplates, any ideas?
[10:38] i mean the link that is in the e-mail.. usually it's something like http://hostname-of-zenoss:8080/.... we can't use the DNS to locate the server for various reasons so it would be more useful if the link was http://10.0.0.2:8080/...
[10:39] malbon: You extended the deviceProperties with zDeviceTemplates at the top of your collector plugin too?
[10:39] cluther: yes, I tried that.
[10:39] zanaga: Add a zopeurl entry to your $ZENHOME/etc/zenactions.conf
[10:40] andyhold: zopeurl http://10.0.0.2:8080
[10:40] oops.. zanaga, not andyhold.
[10:40] ah.. doing that right now
[10:40] malbon: So your getattr is always coming back with ""?
[10:41] zanaga: Restart zenactions when done.
[10:41] cluther: None is what is printed on my debug line
[10:41] cluther: http://pastebin.ca/1187481
[10:41] * rbw has quit ("Lmnar")
[10:41] <__adytum-bot__> Title: general pastebin - Unnamed - post number 1187481 (at pastebin.ca)
[10:41] cluther: a dir on device, and str of deviceProperties
[10:43] curses, i still have some HelloWorld objects haunting my installation..
[10:44] cluther: no rush on it, ping me back later today or tomorrow
[10:46] malbon: How has JMX been working for you in 2.2.3?
[10:47] mrayzenoss: Haydn has been doing most of the testing. I think the attributes have been working ok.
[10:47] malbon: Yeah.. I have your mail on it. I know I've tested this due to a previous thread we've had.
[10:47] malbon: well jplouis has been doing most of that work, so here's your chance to complain
[10:48] mrayzenoss: there is the issue of the methods not returning the right values, well values at all.
[10:48] cluther: yes, I would like to get this sorted but it'd probably be better to take it off-line. I am trying Haydn's hack to fix my itch.
[10:48] for jmx operations or for attribute calls?
[10:48] jplouis: operations
[10:49] with or without parameters?
[10:49] jplouis: I belive that attribute collection and graphing is working.
[10:49] from my understanding with parameters
[10:49] malbon: hold on...
[10:49] sadly he isn't here to ask.
[10:50] cluther: do you want an egg or zipfile of my code?
[10:51] malbon: http://dev.zenoss.com/trac/ticket/3573
[10:52] malbon: zip is good
[10:53] kgoedtel: Looks like there's a catalog indexing problem.
[10:53] kgoedtel: For the time being, do a dmd.Devices.reIndex() in zendmd, then a commit()
[10:53] so 2.2.4 is going to drop soon?
[10:53] kgoedtel: This will fix the existing reports.
[10:53] malbon: Yeah, 2.2.4 is in QA right now
[10:53] cluther: great.. zopeurl does exactly what i wanted. Thanks
[10:57] mrayzenoss: cool, I'll look forward to the upgrade fun
[10:57] jplouis: have you scaled this to many hosts?
[10:57] 2.3 should be in beta next month and is slated for October
[10:58] 2.3 has some enhancements to deal with a lot of hosts
[10:58] sounds good. what leet features does it include?
[10:58] cluther: worked like a charm!
[10:59] malbon: I meant jmx has some features for more hosts
[10:59] I must admit we suffer from slowdowns especially on high zeneventlog insertion rates, also on models with software inventories.
[10:59] malbon: the number of connections to the jmx servers should be greatly reduced, mainly one connection per collection cycle
[11:00] jplouis: yeah good. We will have it on over 70 servers soon so it'll get a good testing.
[11:00] 2.3 will have JP's JMX improvements, WMI improvements (Vista & Server 2008) and more VMware stuff
[11:00] malbon: zenjmx in 2.3 will also be better about sending events on error conditions (bad mbean name etc..)
[11:00] jplouis: cool, that's actually useful.
[11:01] jplouis: a modeller framework?
[11:01] malbon: no, not this go around
[11:01] jplouis: shame, I was hoping.
[11:02] jplouis: talking of this let me know when you get to that stage, I'd be interested in jython integration for modelling.
[11:02] 2.3 is more about stabilizing and improving what we have, more features will come around in the next release
[11:03] yeah, that's good, there has been some good work done in the 2.2 series to that goal too.
[11:04] JP and I are going to take off pretty soon, thanks for everyone's questions. The log will be up soon and we'll have another session scheduled for 2 weeks.
[11:04] malbon: if you have a test system you can use the trunk version of zenjmx for load testing.
[11:05] jplouis: I think haydn is doing that on his test system, so I'll check on where he is.
[11:05] And thanks to cluther for answering all the questions for us
[11:05] malbon: ok, the newest zenjmx is somewhat different... it actually is a python daemon and an small java rpc server
[11:06] bye
[11:06] jplouis: ok, nice.
[11:06] * jplouis has quit ()
[11:06] thanks zenoss folks.
[11:07] later
[10:00] * Topic for #zenoss is: Zenoss Development will be here Thursday August 28 at 11am EST
[10:00] * Topic for #zenoss set by mrayzenoss at Thu Aug 14 16:06:33 2008
[10:01] Good morning.
[10:01] * jplouis (n=jplouis@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:01] hey matt
[10:02] jplouis and I are ready for questions, other Zenossians are lurking
[10:03] * _chris_ (n=chris@75-139-111-215.dhcp.mant.nc.charter.com) has joined #zenoss
[10:03] oh, I forgot the devs are here today..
[10:03] just one
[10:03] so far
[10:03] im guessing there was no progress on the whole RRD gaps/zenperfsnmp problems..
[10:03] so hopefully when I transition to enterprise this problem will go away..
[10:03] well, the same people will be working on them
[10:04] Chet and EAD were on that
[10:04] * bedwards (n=bedwards@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:04] * cluther (n=cluther@ip72-192-244-133.dc.dc.cox.net) has joined #zenoss
[10:04] speaking of the devil
[10:05] Hey Chet, jb was just asking about http://dev.zenoss.com/trac/ticket/3536
[10:05] wait.. I wasn't the devil, was I?
[10:06]
[10:08] * Erik-Zenoss (n=edahl@74.93.194.82) has joined #zenoss
[10:09] * Erik-Zenoss has quit (Client Quit)
[10:09] * Erik-Zenoss (n=edahl@74.93.194.82) has joined #zenoss
[10:10] * bwhichard (n=bwhichar@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:10] * bootay (n=bootay@rrcs-97-77-9-2.sw.biz.rr.com) has joined #zenoss
[10:10] * Erik-Zenoss has quit (Client Quit)
[10:11] any questions?
[10:11] seems pretty quiet today
[10:11] I have systems being shown in reports when using systems organizer in which they no longer exist in the organizer the particular report is using.
[10:12] mrayze: is there any way i can have the status of zencommand based checks be shown in the device status screen when they are not in error condition? we would like the reassuring green pill
[10:12] The device was removed from one organizer and placed in another, but it still is 'residually' reporting for the previous organizers
[10:12] this is 2.2.3
[10:12] teletype telexicon
[10:13] * dundel has quit (Read error: 113 (No route to host))
[10:15] I have been seeing strange errors in zenactions that seem to cause it to time out quite often when I have email rules set with counts.. Ver 2.2.3 x64
[10:15] kgoedtel: can you open a ticket with steps to reproduce?
[10:15] * paulh_ (n=paulh@74.93.194.82) has joined #zenoss
[10:16] I haven't been able to reproduce at will unfortunately
[10:16] i've viewed the organizer via zope manage to make sure i wasn't insane
[10:17] teletype: The individual device will show "green" on the device page, but you want it on the dashboard?
[10:17] teletype telexicon
[10:17] nope, on the device page status tab. zencommand based checks are only showing up in the status display if they are in an error state.
[10:17] teletype telexicon
[10:18] teletype: that sounds like a real bug, why don't you open a ticket
[10:18] kgoedtel: any steps and which reports would be helpful to track it down.
[10:18] ah, really? so i'm not crazy then. will do, thanks
[10:19] basically all i did was remove a device from one organizer under systems to another
[10:19] * cgibbons (n=cgibbons@75.148.215.110) has joined #zenoss
[10:19] i can provide screenshots
[10:19] which would probably illustrate the problem best
[10:19] teletype telexicon
[10:20] teletype: we'll verify the ticket. You're just running zencommand and nothing else, and it stays grey?
[10:20] Guyverix: didn't you have a forum thread for that?
[10:20] actually I opened a bug.
[10:21] was kinda curious if it was me, or a real bug
[10:21] i'm trying to use some custom nagios scripts via zencommand over ssh. the check works correctly, but if i look at the device status page, it does not show up at all. if i cause the script to return an error state, then the component is listed with the appropriate color pill under 'Other' category
[10:21] if i then cause the script to return an OK status, the component will no longer be listed
[10:23] Guyverix: ticket #?
[10:23] let me dbl check
[10:24] kgoedtel: screenshots and the report being run would be helpful
[10:24] lo matt el al.
[10:25] kgoedtel: This organizer/reporting behavior is expected. The organizers that an event is associated with are set at event time, not reporting time. So if an event is created for a device when it is in the /Mail system, then the device gets moved to the /Database system the event will remain a /Mail event.
[10:25] are there any examples of using SnmpPlugin.deviceProperties
[10:25] ?
[10:25] malbon: $ZENHOME/Products/DataCollector/plugins/zenoss/snmp/HRFileSystemMap.py
[10:26] bug 3593
[10:26] hrmf
[10:26] Guyverix: I just found it
[10:27] this report is a performance report
[10:27] Guyverix: and zenactions hangs and that's the last thing in the log?
[10:27] not an event report
[10:27] yes
[10:27] cluther: thanks.
[10:28] when I restart it, it says that it finished the action after usually over 1000 seconds
[10:28] Guyverix: I'll raise the priority, that doesn't sound like a medium bug
[10:28] kgoedtel: Which report(s)?
[10:28] ok, I pasted in what I think is the sql command being used, but I dont know if the h. is a place holder or not
[10:29] a custom one i created in multi-graph
[10:30] kgoedtel: ah.. ok.
[10:30] Guyverix: that ticket will get looked at, it looks like a good candidate to get fixed for 2.3
[10:31] i am getting some screen, have em in a bit
[10:31] cool!
[10:31] BTW, is there anything in the works to integrate Zenoss into external ticketing systems?
[10:32] Guyverix: Enterprise has a Remedy plugin
[10:32] Guyverix: I have done this using email and our ticketing system.
[10:32] There is an AlarmPoint ZenPack coming someday
[10:33] Ahh, OK.. So Remedy, via enterprise, and RT or the like via email..
[10:33] yeah, I'd like a good RT howto
[10:33] does the enterprise one for Remedy show the ticket that is associated with the device?
[10:34] I don't know the specifics, cluther?
[10:34] inside Zenoss itself, I mena
[10:34] Guyverix: Nope.. just Zenoss details in Remedy.
[10:35] hmm, ok
[10:35] kgoedtel: I can definitely reproduce that in the multi-graph reports.
[10:35] one thing about integrating them is that the zenpop doesn't have a mechanism to redirect to 'default' device for sorting out returns from the ticketing system of choice.
[10:36] is it possible to customize the link that comes with the email alert. It would work better for us if the url pointed to the IP of the host
[10:37] ${here\manageIp} ?
[10:37] oh great!
[10:38] cluther: I tried the getattr, however I cannot get a valid return for zDeviceTemplates, any ideas?
[10:38] i mean the link that is in the e-mail.. usually it's something like http://hostname-of-zenoss:8080/.... we can't use the DNS to locate the server for various reasons so it would be more useful if the link was http://10.0.0.2:8080/...
[10:39] malbon: You extended the deviceProperties with zDeviceTemplates at the top of your collector plugin too?
[10:39] cluther: yes, I tried that.
[10:39] zanaga: Add a zopeurl entry to your $ZENHOME/etc/zenactions.conf
[10:40] andyhold: zopeurl http://10.0.0.2:8080
[10:40] oops.. zanaga, not andyhold.
[10:40] ah.. doing that right now
[10:40] malbon: So your getattr is always coming back with ""?
[10:41] zanaga: Restart zenactions when done.
[10:41] cluther: None is what is printed on my debug line
[10:41] cluther: http://pastebin.ca/1187481
[10:41] * rbw has quit ("Lmnar")
[10:41] <__adytum-bot__> Title: general pastebin - Unnamed - post number 1187481 (at pastebin.ca)
[10:41] cluther: a dir on device, and str of deviceProperties
[10:43] curses, i still have some HelloWorld objects haunting my installation..
[10:44] cluther: no rush on it, ping me back later today or tomorrow
[10:46] malbon: How has JMX been working for you in 2.2.3?
[10:47] mrayzenoss: Haydn has been doing most of the testing. I think the attributes have been working ok.
[10:47] malbon: Yeah.. I have your mail on it. I know I've tested this due to a previous thread we've had.
[10:47] malbon: well jplouis has been doing most of that work, so here's your chance to complain
[10:48] mrayzenoss: there is the issue of the methods not returning the right values, well values at all.
[10:48] cluther: yes, I would like to get this sorted but it'd probably be better to take it off-line. I am trying Haydn's hack to fix my itch.
[10:48] for jmx operations or for attribute calls?
[10:48] jplouis: operations
[10:49] with or without parameters?
[10:49] jplouis: I belive that attribute collection and graphing is working.
[10:49] from my understanding with parameters
[10:49] malbon: hold on...
[10:49] sadly he isn't here to ask.
[10:50] cluther: do you want an egg or zipfile of my code?
[10:51] malbon: http://dev.zenoss.com/trac/ticket/3573
[10:52] malbon: zip is good
[10:53] kgoedtel: Looks like there's a catalog indexing problem.
[10:53] kgoedtel: For the time being, do a dmd.Devices.reIndex() in zendmd, then a commit()
[10:53] so 2.2.4 is going to drop soon?
[10:53] kgoedtel: This will fix the existing reports.
[10:53] malbon: Yeah, 2.2.4 is in QA right now
[10:53] cluther: great.. zopeurl does exactly what i wanted. Thanks
[10:57] mrayzenoss: cool, I'll look forward to the upgrade fun
[10:57] jplouis: have you scaled this to many hosts?
[10:57] 2.3 should be in beta next month and is slated for October
[10:58] 2.3 has some enhancements to deal with a lot of hosts
[10:58] sounds good. what leet features does it include?
[10:58] cluther: worked like a charm!
[10:59] malbon: I meant jmx has some features for more hosts
[10:59] I must admit we suffer from slowdowns especially on high zeneventlog insertion rates, also on models with software inventories.
[10:59] malbon: the number of connections to the jmx servers should be greatly reduced, mainly one connection per collection cycle
[11:00] jplouis: yeah good. We will have it on over 70 servers soon so it'll get a good testing.
[11:00] 2.3 will have JP's JMX improvements, WMI improvements (Vista & Server 2008) and more VMware stuff
[11:00] malbon: zenjmx in 2.3 will also be better about sending events on error conditions (bad mbean name etc..)
[11:00] jplouis: cool, that's actually useful.
[11:01] jplouis: a modeller framework?
[11:01] malbon: no, not this go around
[11:01] jplouis: shame, I was hoping.
[11:02] jplouis: talking of this let me know when you get to that stage, I'd be interested in jython integration for modelling.
[11:02] 2.3 is more about stabilizing and improving what we have, more features will come around in the next release
[11:03] yeah, that's good, there has been some good work done in the 2.2 series to that goal too.
[11:04] JP and I are going to take off pretty soon, thanks for everyone's questions. The log will be up soon and we'll have another session scheduled for 2 weeks.
[11:04] malbon: if you have a test system you can use the trunk version of zenjmx for load testing.
[11:05] jplouis: I think haydn is doing that on his test system, so I'll check on where he is.
[11:05] And thanks to cluther for answering all the questions for us
[11:05] malbon: ok, the newest zenjmx is somewhat different... it actually is a python daemon and an small java rpc server
[11:06] bye
[11:06] jplouis: ok, nice.
[11:06] * jplouis has quit ()
[11:06] thanks zenoss folks.
[11:07] later