Menu
▾
▴
webadmin-list
|
From: Joaquim H. <jo...@we...> - 2006-05-18 21:07:20
|
I have a "master webmin"-server, that does little else apart from being just that. It runs quite a few monitors and collects data from other webmin servers and webmin monitors. On this server, I monitor a server we call "teacher". Quite frequently, but not according to any pattern I've been able to find, one or two of the monitors will report that "webmin has gone down", and then three minutes later, it comes back up. The exact messages are: "Monitor on teacher.domain.com for 'Postfix, teacher' has detected that Webmin is down at Thu May 18 18:21:01 2006" and then "Monitor on teacher.domain.com for 'Postfix, teacher' has detected that the service has gone back up at Thu May 18 18:24:01 2006 The monitors are configured as such: Postfix - failures before reporting 2 - run on host teacher.domain.com - check on schedule? yes, and report on status changes Apache - failures before reporting 2 - run on host teacher.domain.com - check on schedule? yes, and report on status changes I have at least 12 more monitors configured almost identically, none of them exhibit this behavior. The scheduled monitoring is set to check every 3 minutes with offset 0 Send email when "when a service changes status" For your records, the services that are being monitored, postfix and apache, have never been down. Good thing I shave my head, or I'd be pulling my hair out right about now.. -joho |
|
From: Jamie C. <jca...@we...> - 2006-05-18 21:59:57
|
On 18/May/2006 16:07 Joaquim Homrighausen wrote .. > > I have a "master webmin"-server, that does little else apart from being > just that. It runs quite a few monitors and collects data from other webmin > servers and webmin monitors. > > On this server, I monitor a server we call "teacher". Quite frequently, > but not according to any pattern I've been able to find, one or two of > the monitors will report that "webmin has gone down", and then three minutes > later, it comes back up. > > > The exact messages are: > > "Monitor on teacher.domain.com for 'Postfix, teacher' has detected that > Webmin is down at Thu May 18 18:21:01 2006" > > and then > > "Monitor on teacher.domain.com for 'Postfix, teacher' has detected that > the service has gone back up at Thu May 18 18:24:01 2006 > > > The monitors are configured as such: > > Postfix > - failures before reporting 2 > - run on host teacher.domain.com > - check on schedule? yes, and report on status changes > Apache > - failures before reporting 2 > - run on host teacher.domain.com > - check on schedule? yes, and report on status changes > > I have at least 12 more monitors configured almost identically, none of > them exhibit this behavior. > > The scheduled monitoring is set to check every 3 minutes with offset 0 > Send email when "when a service changes status" > > > For your records, the services that are being monitored, postfix and apache, > have never been down. > > > Good thing I shave my head, or I'd be pulling my hair out right about now.. So you have multiple remote monitors for the teacher server, but only one is failing like this? If so, is there any firewall between the two machines that could be blocking ports 10000-10100 ? - Jamie |
|
From: Joaquim H. <jo...@we...> - 2006-05-18 22:09:35
|
>So you have multiple remote monitors for the teacher server, but only one is failing >like this? If so, is there any firewall between the two machines that could be >blocking ports 10000-10100 ? I've changed the port to 1080, and it's being allowed through the firewall. I'm also using an IP Access Control list in Webmin on the teacher server, and have set the firewall's IP-address and the "webmin master" machine's IP-address as allowed addresses. What's annoying is that I don't get "complete failures" and I don't get them "all the time". That's the part that's driving me nuts :-) -joho |
|
From: Jamie C. <jca...@we...> - 2006-05-18 23:24:43
|
On 18/May/2006 17:09 Joaquim Homrighausen wrote .. > > >So you have multiple remote monitors for the teacher server, but only > one is failing > >like this? If so, is there any firewall between the two machines that > could be > >blocking ports 10000-10100 ? > > I've changed the port to 1080, and it's being allowed through the firewall. > I'm also using an IP Access Control list in Webmin on the teacher server, > and have set the firewall's IP-address and the "webmin master" machine's > IP-address as allowed addresses. Make sure you allow a few TCP ports above 1080 (like up to 1100) as well, as Webmin uses them for server-to-server communication .. and blocking them could cause just this kind of periodic failures. > What's annoying is that I don't get "complete failures" and I don't get > them "all the time". That's the part that's driving me nuts :-) Yeah, that is the worst kind of problem to debug! - Jamie |
|
From: Joaquim H. <jo...@we...> - 2006-05-18 23:44:58
|
>Make sure you allow a few TCP ports above 1080 (like up to 1100) as well, as >Webmin uses them for server-to-server communication .. and blocking them >could cause just this kind of periodic failures. OK; if both machines are behind firewalls, which firewall do I need to open for what? Do I need to open both for an inbound range of 1080-1100, or is it teacher that opens the channel back to the master webmin on a different port? Or does the webmin master open both sockets, like a passive FTP connection? -joho |
|
From: Jamie C. <jca...@we...> - 2006-05-19 00:30:03
|
On 18/May/2006 18:44 Joaquim Homrighausen wrote .. > > >Make sure you allow a few TCP ports above 1080 (like up to 1100) as well, > as > >Webmin uses them for server-to-server communication .. and blocking them > >could cause just this kind of periodic failures. > > OK; if both machines are behind firewalls, which firewall do I need to > open for what? > > Do I need to open both for an inbound range of 1080-1100, or is it teacher > that opens the channel back to the master webmin on a different port? Or > does the webmin master open both sockets, like a passive FTP connection? Basically, the 'master' server needs to be able to connect to the 'slave' on ports 1080-1100. The connections are always only one-way. - Jamie |
|
From: Joaquim H. <jo...@we...> - 2006-05-19 12:46:39
|
>Basically, the 'master' server needs to be able to connect to the 'slave' >on ports 1080-1100. The connections are always only one-way. Well, I opened 1080-1100 and I'm still getting this error :-( What's weird is that it's only the POSTFIX monitor that is doing it. I have one diskspace monitor, one Apache monitor and one monitor that checks for a java virtual machine running. Only the POSTFIX monitor fails. And PF isn't down. -joho |
|
From: Jamie C. <jca...@we...> - 2006-05-19 16:38:56
|
On 19/May/2006 07:46 Joaquim Homrighausen wrote .. > > >Basically, the 'master' server needs to be able to connect to the 'slave' > >on ports 1080-1100. The connections are always only one-way. > > Well, I opened 1080-1100 and I'm still getting this error :-( > > What's weird is that it's only the POSTFIX monitor that is doing it. I > have one diskspace monitor, one Apache monitor and one monitor that checks > for a java virtual machine running. Only the POSTFIX monitor fails. And > PF isn't down. Do you know which order the monitors are executed in? I am wondering if it happens because the Postfix monitor is last. Or perhaps because the monitor before it is failing in some way .. - Jamie |
|
From: Joaquim H. <jo...@we...> - 2006-05-19 16:49:22
|
>Do you know which order the monitors are executed in? I am wondering if it >happens because the Postfix monitor is last. Or perhaps because the monitor >before it is failing in some way .. Is there a way I can find out in which order they're executed? I assume since you ask that it's not in the order of appearance.. ? -joho |
|
From: Jamie C. <jca...@we...> - 2006-05-19 16:58:38
|
On 19/May/2006 11:49 Joaquim Homrighausen wrote .. > > >Do you know which order the monitors are executed in? I am wondering > if it > >happens because the Postfix monitor is last. Or perhaps because the monitor > >before it is failing in some way .. > > Is there a way I can find out in which order they're executed? > I assume since you ask that it's not in the order of appearance.. ? On the Module Config page, change the 'Display monitors sorted by' to 'Order created'. They will then be shown in the same order that they are actually run. - Jamie |
|
From: Joaquim H. <jo...@we...> - 2006-05-19 21:39:36
|
> Do you know which order the monitors are executed in? I am wondering if it > happens because the Postfix monitor is last. Or perhaps because the monitor > before it is failing in some way .. The one that is failing is fourth from last. The one before and the one after never seem to fail (and look OK now, as we "speak", and the other one just gave a failure message again). -joho |
|
From: Fred B. <ba...@ae...> - 2006-05-20 17:25:39
|
Jamie, I set up a new machine yesterday and installed the latest version of Webmin on it. One of the first things I did was to try to install some 3rd party modules through the Webmin Modules interface. Previously, when selecting "Third party module from" and clicking on the "Select" button, a window would open with a list of 3rd party modules to choose from. Now that window is empty except for the header. I tried this on my other machines with the same result. Is this a new bug or a new feature? ;-) This appears to have started with version 1.270. Fred Bacon |
|
From: Jamie C. <jca...@we...> - 2006-05-21 05:56:31
|
On 20/May/2006 12:24 Fred Bacon wrote .. > Jamie, > > I set up a new machine yesterday and installed the latest version of > Webmin on it. One of the first things I did was to try to install some > 3rd party modules through the Webmin Modules interface. Previously, > when selecting "Third party module from" and clicking on the "Select" > button, a window would open with a list of 3rd party modules to choose > from. Now that window is empty except for the header. I tried this on > my other machines with the same result. Is this a new bug or a new > feature? ;-) This appears to have started with version 1.270. This was actually caused by a problem on the www.webmin.com server, in the CGI that lists third-party modules. Thanks for pointing it out - it is now fixed! - Jamie |
