October 2010 - Xymon - Xymon mailman web

Memory report faulty
by everett.vernon＠gmail.com 11 Oct '10

11 Oct '10

Hi all Another contract, another emial address. :-) Hope you all had a good festive season. Quick question. Was there ever a resolution to this issue? http://www.hswn.dk/hobbiton/2009/03/msg00036.html Searched the archive, and couldn't find one. To recap, this was where memory reports something bizzare, like this > Memory Used Total Percentage > red Physical 4294961768M 8192M 4294967229% > green Swap 1600M 8191M 19% Regards Vernon

4 6

HTML e-mails to this list are silently dropped?
by spa＠syntec.co.uk 09 Oct '10

09 Oct '10

I was trying to send e-mail to this list and they were being silently dropped rather than published. I changed my e-mail format from HTML to Plain Text and - hey, presto - my e-mail is not dropped! Trying this e-mail in HTML format first to see if it reproducible on a simpler message (without tabs and numbering)... If this is published, it points to tabs or numbering being disliked by the mail program. If it is not published, further evidence that HTML containing e-mails are dropped... Regards, SebA

4 3

problem with subpagecolums
by akshar.bhosale＠gmail.com 09 Oct '10

09 Oct '10

hello Henrik/Xymon Team, Is subpagecolumns issue resolved in 4.3 version?please help us as we are praising xymon and encouraging other sysadmin teams to use the same. your kind help will be appreciated. ---------- Forwarded message ---------- From: akshar bhosale <akshar.bhosale at gmail.com> Date: Mon, Oct 4, 2010 at 12:01 AM Subject: subcolumns and display issue To: xymon at xymon.com Hi, we have xymon4.2.2 and rhel 5.2 os. we have a server and 100 clients. We have 3 sub clusters and above bb-host file. Now our problem is that we have restriction on subcolumns argument to BBGENOPTS as this is common for all three subcluster pages..so if we want to add one extra test in one of the subcluster, we need to increase subcolumns aruments..e.g.all subclusters have 8 tests, so subcolumns=8, now if we want to configure one more test in subcluster2 only, how we should do that? because if we do subcolumns=9, all our display page will be disturbed as there are 8 tests for subcluster1 and 3 and 9 tests for subcluster2. So how to add or remove any service without disturbing the main display our hostfile is like: page rack001 Compute cluster title Compute Nodes subpage rack900 Row1-Rack6 conn group-only conn 172.29.1.49 node49.mgmt.acc.in # 172.29.1.50 node50.mgmt.acc.in # 172.29.1.51 node51.mgmt.acc.in # 172.29.1.52 node52.mgmt.acc.in # 172.29.1.53 node53.mgmt.acc.in # 172.29.1.54 node54.mgmt.acc.in # subpage rack901 cpu group-only cpu 0.0.0.0 node49.mgmt.acc.in # noconn 0.0.0.0 node50.mgmt.acc.in # noconn 0.0.0.0 node51.mgmt.acc.in # noconn 0.0.0.0 node52.mgmt.acc.in # noconn 0.0.0.0 node53.mgmt.acc.in # noconn 0.0.0.0 node54.mgmt.acc.in # noconn subpage rack904 ssh group-only ssh 0.0.0.0 node49.mgmt.acc.in # noconn 0.0.0.0 node50.mgmt.acc.in # noconn 0.0.0.0 node51.mgmt.acc.in # noconn 0.0.0.0 node52.mgmt.acc.in # noconn 0.0.0.0 node53.mgmt.acc.in # noconn 0.0.0.0 node54.mgmt.acc.in # noconn subpage rack904 ssh group-only ssh 0.0.0.0 node49.mgmt.acc.in # noconn 0.0.0.0 node50.mgmt.acc.in # noconn 0.0.0.0 node51.mgmt.acc.in # noconn 0.0.0.0 node52.mgmt.acc.in # noconn 0.0.0.0 node53.mgmt.acc.in # noconn 0.0.0.0 node54.mgmt.acc.in # noconn ....and so on ? Kindly help

4 5

lib/url.c patch for '/http' malformed url?
by viorel.tabara＠infrabit.net 08 Oct '10

08 Oct '10

As discussed at http://www.xymon.com/archive/2006/07/msg00515.html: > On Mon, Jul 31, 2006 at 06:50:19PM +0200, Nicolas wrote: > > It doesn t like the "%3a" strings. > > Apparently not. OK, so this will require a code change. For a > quick fix, you can edit the lib/url.c file in your Hobbit source. > Around line 560-570 or so you'll see these lines: > > if (poststart) getescapestring(poststart, &bburl->postdata, NULL); > if (expstart) getescapestring(expstart, &bburl->expdata, NULL); > > p = strstr(urlstart, "/http"); <--- find this line > if (p) { > proxystart = urlstart; > urlstart = (p+1); > *p = '\0'; > } > > parse_url(urlstart, bburl->desturl); > > Replace the line I've marked with > > p = NULL; > > That will simply disable the proxy url function, and your original URL > should work as intended. > > > Regards, > Henrik > Hi Henrik, Obviously it's not the ":" and I wish I would've searched the mailing list earlier ;) # /usr/libexec/hobbit/server/bin/bbtest-net 2010-10-08 14:31:38 Malformed URL - no 'scheme:' in 'http-proxy-tcp' 2010-10-08 14:31:38 Invalid URL for http test - ignored: cont=postgresql;http://my.athabascau.ca/http-proxy-tcp;0 Changed that to: cont=postgresql;http://my.athabascau.ca/tohack;0 and it works well. Do you have any plans to release an update? Thanks, Viorel

1 0

How do *you* handle Xymon alerts when you are on vacation (holiday in British English)?
by robert.herron＠gmail.com 08 Oct '10

08 Oct '10

Vacation / holiday? What's that? But seriously. Since my Xymon server is a fully capable Linux box, my hobbit-alerts.cfg sends emailto aliases defined in the /etc/aliases. If/when someone wants messages disabled for vacation, I just change the alias to send the message somewhere else. I like this method as it prevents needing to make changes to my fairly complex/convoluted hobbit-alerts.cfg file and an errant update to the hobbit-alerts.cfg could cause alerting to not work. I don't use a native SMS gateway and instead use the cell provider's email to text/SMS gateway. Robert Herron robert.herron at gmail.com On Fri, Oct 8, 2010 at 2:30 PM, <xymon-digest-help at xymon.com> wrote: > ---------- Forwarded message ---------- > From: "SebA" <spa at syntec.co.uk> > To: <xymon at xymon.com> > Date: Fri, 8 Oct 2010 19:30:18 +0100 > Subject: How do *you* handle Xymon alerts when you are on vacation (holiday > in British English)? > Hi all, > > I'm interested in the different ways that people handle alerts when they > are > on > vacation (holiday in British English)... I'm (still!) migrating from Big > Brother BTF (still mostly use that for alerts, Xymon for the rest), and if > I > > went on vacation, I'd just put 1 line at the bottom of bb-alerts excluding > me > from all alerts (as they get SMSed to my mobile) and someone else would get > them (as I was never the only one to get them anyway). But you can't do > that > in Hobbit / Xymon (as I posted about last year). > >

1 0

HTML e-mails to this list are silently dropped?
by dwdaniel＠vcu.edu 08 Oct '10

08 Oct '10

Return Receipt Your [xymon] HTML e-mails to this list are silently dropped? document: was received by: at: 10/08/2010 15:01:16

1 0

How do *you* handle Xymon alerts when you are on vacation (holiday in British English)?
by spa＠syntec.co.uk 08 Oct '10

08 Oct '10

Hi all, I'm interested in the different ways that people handle alerts when they are on vacation (holiday in British English)... I'm (still!) migrating from Big Brother BTF (still mostly use that for alerts, Xymon for the rest), and if I went on vacation, I'd just put 1 line at the bottom of bb-alerts excluding me from all alerts (as they get SMSed to my mobile) and someone else would get them (as I was never the only one to get them anyway). But you can't do that in Hobbit / Xymon (as I posted about last year). So how do you do it, or what do you recommend for a network containing a variety of different machines, routers, O/Ses and services, where the primary, secondary and tertiary responsibilities for each service may be different people depending on the service - in the case where one of those people goes on vacation? A few ways I can think of doing it using features I know already exist: 1. have a distinct e-mail address for each person that receives alerts, that should only receive alerts and no other e-mails, and then redirecting that address to another person when they go on holiday and then redirecting back to the original person when they come back (has the downsides of meaning that the other person gets all their alerts, even if they are not the most suitable person or getting those alerts anyway and so ending up with 2 copies of them - particularly annoying if they are SMSes like we use; another downside is remembering to put it back when you get back from holiday) 2. going through all the services for which the person going on vacation receives alerts, and changing them so that they go to someone else... Has the downsides of the tedium involved in doing it and reverting it... Might be easier if you have a special macro for each set of changes, e.g. PersonA->PersonB, PersonA->PersonC and PersonA->Person C is 3 macros... It is easy to change the macro when you get back (if you remember). 3. following on from #2, I suppose instead of defining macros for each recipient (which is already one layer of abstraction better then the 'default'), you could define a macro for each set of primary, secondary and tertiary recipients (although that could make quite a few permutations in a large office). Then you just have to change each of those macros when someone goes on holiday, and then again when they get back. 4. A custom script, possibly with a database and putting all the intelligence in there (but that probably doesn't help others here, unless the script is publicly available). 5. Er, and passing mobile (cell) phones back and forth - leaving your work mobile with someone else when you have the day off - but that only works if you have a work mobile! And then there are the ways using features that have not so far been built: 1. being able to define vacations (I know public holidays has been added to a revent version) for recipients, when they start and when they finish... being able to set a VACATIONRECIPIENT who gets the alert if the first recipient is on vacation (they will need to not get it if they are listed as the 2nd recipient too) 2. as above, but without the VACATIONRECIPIENT. Instead if one of the recipients is defined as on vacation, all the others automatically shift up in terms of any DURATION> required before sending them an alert 3. having the exclude feature that BB had, but, althought that was easy to do before going on holiday, it was a poor solution anyway in terms of handling of alerts 4. some other way... I'm not going to spend too much time thinking of them now when there may be a good solution already! ;) Kind regards, SebA No virus found in this outgoing message. Checked by AVG - www.avg.com Version: 8.5.448 / Virus Database: 271.1.1/3149 - Release Date: 10/07/10 18:34:00

1 1

test... ignore if this reaches the list
by dwdaniel＠vcu.edu 08 Oct '10

08 Oct '10

Return Receipt Your [xymon] test... ignore if this reaches the list document: was received by: at: 10/08/2010 13:17:56

1 0

test... ignore if this reaches the list
by spa＠syntec.co.uk 08 Oct '10

08 Oct '10

I've been trying to send a message to this list for a few weeks now and it never comes back to me as a subscriber or goes into the hobbit archives so I guess it's being silently deleted somewhere (no reject message either), so a little test with a simple message to see whether it was something in my message that was not liked, or whether it is (just) due to move from hobbit to xymon or something else screwing with the subscriptions... Kind regards, SebA

1 0

hobbit-alerts.cfg question
by colin.coe＠gmail.com 08 Oct '10

08 Oct '10

Hi all The alerting is starting to take shape but I've a question regarding how the alerting works. If I have a stanza similar to the following, how is it evaluated? Once for all hosts, or for one host at a time? --- HOST=%.* # Proliant tests MAIL sms at somecompany.com SERVICE=proliant FORMAT=SMS REPEAT=1440m MAIL sms at somecompany.com SERVICE=proliant FORMAT=SMS RECOVERED # conn where status is RED MAIL sms at somecompany.com COLOR=red SERVICE=conn EXPAGE=dev REPEAT=1440m MAIL sms at somecompany.com COLOR=red SERVICE=conn EXPAGE=dev RECOVERED # conn where status is RED (dev/test) MAIL email at somecompany.com COLOR=red SERVICE=conn PAGE=dev REPEAT=1440m MAIL email at somecompany.com COLOR=red SERVICE=conn PAGE=dev RECOVERED # cpu,disk,memory where status is RED MAIL sms at somecompany.com COLOR=red SERVICE=cpu,disk,memory EXPAGE=dev REPEAT=1440m MAIL sms at somecompany.com COLOR=red SERVICE=cpu,disk,memory EXPAGE=dev RECOVERED # Dev servers MAIL email at somecompany.com COLOR=red SERVICE=cpu,disk,memory PAGE=dev REPEAT=1440m MAIL email at somecompany.com COLOR=red SERVICE=cpu,disk,memory PAGE=dev RECOVERED # Non-dev status YELLOW MAIL email at somecompany.com COLOR=yellow SERVICE=cpu,disk,memory REPEAT=1440m DURATION>30m MAIL email at somecompany.com COLOR=yellow SERVICE=cpu,disk,memory RECOVERED --- Also, I've noticed that when a fault occurs I get two emails (or sms') and another when the fault is rectified. I'm thinking this is because of the 'RECOVERED' line but i thought this would only trigger when the fault goes. Have I misunderstood? Thanks CC -- RHCE#805007969328369

3 3