monami-users Mailing List for MonAMI - your friendly monitoring daemon (Page 3)
Status: Alpha
Brought to you by:
paulmillar
You can subscribe to this list here.
2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
(2) |
Sep
|
Oct
|
Nov
(2) |
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2008 |
Jan
|
Feb
(8) |
Mar
(13) |
Apr
(16) |
May
(13) |
Jun
(9) |
Jul
(11) |
Aug
(3) |
Sep
(4) |
Oct
(2) |
Nov
(6) |
Dec
(6) |
2009 |
Jan
|
Feb
|
Mar
(6) |
Apr
|
May
(9) |
Jun
(1) |
Jul
(10) |
Aug
(6) |
Sep
|
Oct
(3) |
Nov
(7) |
Dec
|
2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
From: Stephen C. <ch...@cs...> - 2008-11-27 09:21:08
|
Paul Millar wrote: > Hi Stephen, > > On Tuesday 25 November 2008 16:49:19 Stephen Childs wrote: >> I just rebuilt monami to get support for my current version of torque >> (torque-2.3.0-2cri.slc4.x86_64 -- can circulate rpms if there's interest). > > That would be useful! One of the problems is trying to build RPMs for all the > different possible platforms. RPMs available at http://grid.ie/distribution/monami/x86_64/SL45 (although as per my other email the torque plugin doesn't seem to work very well!). As for the MAUI password, I just got it from the person who built the MAUI RPMs! Stephen -- Dr. Stephen Childs, Research Fellow, EGEE Project, phone: +353-1-8961797 Computer Architecture Group, email: Stephen.Childs @ cs.tcd.ie Trinity College Dublin, Ireland web: http://www.cs.tcd.ie/Stephen.Childs |
From: Paul M. <p.m...@ph...> - 2008-11-27 01:04:43
|
Hi Stephen, On Tuesday 25 November 2008 16:49:19 Stephen Childs wrote: > I just rebuilt monami to get support for my current version of torque > (torque-2.3.0-2cri.slc4.x86_64 -- can circulate rpms if there's interest). That would be useful! One of the problems is trying to build RPMs for all the different possible platforms. > It fixed torque output ok but broke maui! I see the following: > > maui> target_instance( maui, 0x0x6a1bc0) > maui> sorry, currently no support for 64-bit ELF file /usr/bin/mclient > maui> no password for maui server, please specify one or install a maui > executable > > Is there any way to make this work? Probably... > what is the problem with 64-bit? So, at the risk of pointing out what you've already read, the general problem is described here: http://monami.sourceforge.net/userguide/ch03s04.html#maui-password Inside the maui plugin is an ELF parser (the format of executables on Linux, Solaris, etc). The parser attempts to extract the "secret" from a maui executable using a combination of the ELF-parsing and some heuristics. The ELF parser was written against the 32-bit ELF spec. Support in ELF for 64-bit platforms is mentioned, but only insofar as stating a single flag should be set. Otherwise, 64-bit platforms was not mentioned at all. AMD have released a spec for the AMD-64 ABI, which includes a definition of the necessary ELF extensions (e.g., 64-bit address space). I've had a quick shifty through it and it doesn't look too bad. So, adding 64-bit support may be possible. I'll have to have a look. Cheers, Paul. |
From: Paul M. <p.m...@ph...> - 2008-11-27 00:45:29
|
Hi Stephen, On Wednesday 26 November 2008 17:27:27 Stephen Childs wrote: > Now I have monami consuming loads of CPU and the torque plugin isn't > working. That sounds like a bug. Which version of MonAMI are you using? There was a similar issue with the torque plugin in MonAMI v0.9 and earlier (triggered by restarting the torque server, iirc) but this should have been fixed with v0.10. If you're not running v0.10, I would recommend upgrading and seeing if that fixes the problem. If you're running v0.10, could you try gathering a thread timing profile? You can do that by sending monamid a sigusr1 signal; the thread profile should appear as info messages (so, monamid -fv should report this on stdout). > Errors are as below. What do they mean? It looks like you have full debugging switched on. This is often helpful, but it means you're seeing things that are not always errors; they're more "useful information". > torque> process_msg() [anonymous sample 1> --> 2 (std) This indicates that the torque plugin has received a message from "anonymous sample 1" target to the torque target's standard port: port 2 (PORT_DATA_REQUEST). That is, the (anonymous) sample is attempting to collect data from the torque plugin. This is expected. > torque> unknown total_jobs from pbs_statserver(). > torque> unknown state_count from pbs_statserver(). > torque> unknown node_ping_rate from pbs_statserver(). > [...] The torque API function pbs_statserver() returns a number of keyword-value pairs describing the current status of the server. The torque plugin has a built-in list of keywords it is interested in and some static information about the metric: what kind of metric it is (string, integer), where in the datatree to place the metric, that sort of thing. MonAMI isn't interested in all the metrics coming from pbs_statserver(), so some of them it will ignore. These messages describe which metrics it is ignoring. (Perhaps we could do better here and silence the ones that are deliberately ignored). HTH, Paul. |
From: Stephen C. <ch...@cs...> - 2008-11-26 16:28:44
|
Now I have monami consuming loads of CPU and the torque plugin isn't working. Errors are as below. What do they mean? torque> process_msg() [anonymous sample 1> --> 2 (std) torque> unknown total_jobs from pbs_statserver(). torque> unknown state_count from pbs_statserver(). torque> unknown node_ping_rate from pbs_statserver(). torque> unknown job_stat_rate from pbs_statserver(). torque> unknown poll_jobs from pbs_statserver(). torque> unknown log_level from pbs_statserver(). torque> unknown mom_job_sync from pbs_statserver(). torque> unknown mail_domain from pbs_statserver(). torque> unknown log_file_max_size from pbs_statserver(). torque> unknown log_file_roll_depth from pbs_statserver(). torque> unknown next_job_number from pbs_statserver(). torque> unknown net_counter from pbs_statserver(). torque> unknown server_name from pbs_statserver(). torque> no structure cache, storing it. -- Dr. Stephen Childs, Research Fellow, EGEE Project, phone: +353-1-8961797 Computer Architecture Group, email: Stephen.Childs @ cs.tcd.ie Trinity College Dublin, Ireland web: http://www.cs.tcd.ie/Stephen.Childs |
From: Stephen C. <ch...@cs...> - 2008-11-25 15:49:25
|
Hi, I just rebuilt monami to get support for my current version of torque (torque-2.3.0-2cri.slc4.x86_64 -- can circulate rpms if there's interest). It fixed torque output ok but broke maui! I see the following: maui> target_instance( maui, 0x0x6a1bc0) maui> sorry, currently no support for 64-bit ELF file /usr/bin/mclient maui> no password for maui server, please specify one or install a maui executable Is there any way to make this work? what is the problem with 64-bit? Stephen -- Dr. Stephen Childs, Research Fellow, EGEE Project, phone: +353-1-8961797 Computer Architecture Group, email: Stephen.Childs @ cs.tcd.ie Trinity College Dublin, Ireland web: http://www.cs.tcd.ie/Stephen.Childs |
From: Paul M. <p.m...@ph...> - 2008-10-22 21:11:18
|
Hi Liam, On Wednesday 22 October 2008 15:23:56 Liam Gretton wrote: > It seems that MonAMI 0.10 doesn't send gmetrics info to Ganglia 3.1.1 > (the latest verson); certainly Ganglia doesn't see anything at all from > it as far as I can tell. The Ganglia 3.1.0 release notes suggest that > the gmond comms protocol has changed significantly, I guess this is why > MonAMI isn't working with it? That's correct. The 3.1-series of Ganglia has a non-backwards compatible change to the data format. Currently, the ganglia plugin only supports the older v3.0.x format. (To be honest, I wasn't sure how widely deployed is the 3.1-series and had delaying working on the new format until someone complained ;-) I'll try to have a look at the new format. > Also it looks like the show-graph-3.0.7.diff file needs updating for 3.1.1: > > patch -p0 < patches/show-graph-3.0.7.diff > > patching file host_view.php > Hunk #1 FAILED at 106. > 1 out of 1 hunk FAILED -- saving rejects to file host_view.php.rej > patching file header.php > Hunk #1 succeeded at 339 (offset 5 lines). > patching file get_context.php > Hunk #1 succeeded at 21 with fuzz 1 (offset 4 lines). > patching file templates/default/header.tpl > Hunk #1 succeeded at 34 (offset 1 line). I'm currently working on improving the mg_frame support for a couple of targets and noticed there was a problem when patching v3.0.5 (the in latest v3.0.x series), so I was planning on fixing this anyway. I noticed that the v3.0.5 web front-end has support for metric groups (albeit with all metrics in a default group). IIRC, metric groups is a 3.1.x-series concept, so it's possible that fixing mg_frame for v3.0.5 may provide a fix for v3.1.x. ... we shall see! Cheers, Paul. |
From: Liam G. <L.G...@lb...> - 2008-10-22 13:24:20
|
Hi, It seems that MonAMI 0.10 doesn't send gmetrics info to Ganglia 3.1.1 (the latest verson); certainly Ganglia doesn't see anything at all from it as far as I can tell. The Ganglia 3.1.0 release notes suggest that the gmond comms protocol has changed significantly, I guess this is why MonAMI isn't working with it? Also it looks like the show-graph-3.0.7.diff file needs updating for 3.1.1: > patch -p0 < patches/show-graph-3.0.7.diff patching file host_view.php Hunk #1 FAILED at 106. 1 out of 1 hunk FAILED -- saving rejects to file host_view.php.rej patching file header.php Hunk #1 succeeded at 339 (offset 5 lines). patching file get_context.php Hunk #1 succeeded at 21 with fuzz 1 (offset 4 lines). patching file templates/default/header.tpl Hunk #1 succeeded at 34 (offset 1 line). -- Liam Gretton L.G...@lb... IT Services http://www.lboro.ac.uk/ Loughborough University Tel: +44 (0)1509 228431 Leicestershire LE11 3TU United Kingdom |
From: Paul M. <p.m...@ph...> - 2008-09-20 22:57:14
|
Hi Tomas, On Tuesday 16 September 2008 12:08:53 Tomas Kouba wrote: > I attach a small patch fixing issues you raised. Many thanks for this. I've applied the new version to CVS along with the patches to the "external" package (getting Ganglia to display the extra data). I made a few very minor adjustments before committing the changes: With the dpm plugin code, I took out the comment-ed out sections of code. I guess they were left in by mistake and have removed them. With the PHP code, you used "t" to pass the token description to graph-dpm.php; unfortunately, this is already used by Ganglia for "tree" (see get_context.php). To avoid any potential confusion, I changed token to "tk". You also seemed to have a duplicate "</td> </tr>\n" inside the token pop-up; although it should be mostly harmless (I think it will generate a blank line within the pop-up table) I removed it for safety. Please let me know if I've broken anything! Cheers, Paul. |
From: Tomas K. <to...@ji...> - 2008-09-16 03:09:13
|
Paul Millar wrote: > Hi Tomas, > > On Thursday 11 September 2008 09:37:52 Tomas Kouba wrote: >> I have done some patching for tokens support. The patch (against current >> cvs) is attached. > > Looks good! I'm afraid I can't look at merging this until next week as I'm > currently at Karlsruhe for the GridKa school. > >> It is a simple query to mysql database and patch for monami ganglia >> externals. > > A few questions: > > It looks like you're reporting two metrics for space reservations: "used" > and "free". Would it be useful to publish a "total" metric of the > reservation? (I think the other plugins tend to err towards publishing too > much information :-) I guess it would be useful (patch attached). > The "used" metric is published as a 64-bit (signed) integer and "free" as > unsigned 32-bit integer. Is this difference intentional? No it was a mistake (patch attached). > > How do people feel about publishing the capacities as an unsigned 32-bit > integer? If I've got my sums right, this would give a maximum value of 4PT. > If that's a little too tight, we could report the capacity metrics in GB. I am not sure about this, because people can have small (MB) pools or tokens dedicated for testing or operations (SAM tests). I attach a small patch fixing issues you raised. -- Tomas Kouba diff -ur monami.orig/MonAMI/src/dpm/getvalues.c monami/MonAMI/src/dpm/getvalues.c --- monami.orig/MonAMI/src/dpm/getvalues.c 2008-01-22 09:27:01.000000000 +0100 +++ monami/MonAMI/src/dpm/getvalues.c 2008-09-16 11:58:18.000000000 +0200 @@ -276,3 +276,86 @@ mysql_free_result(results); } + +void dpm_get_tokens(dpm_handle_t *h, void *root) +{ + MYSQL_RES *results; + MYSQL_ROW row; + void *tokens_root_branch; + void *result_tokens_used; + void *result_tokens_free; + void *result_tokens_total; + unsigned int total, used, free; + datatree_value_t used_val; + datatree_value_t total_val; + + if( mysql_select_db( h->conn, "dpm_db")){ + dpm_print_error( h->conn, "Selection of the dpm_db failed"); + return; + } + /* perform query */ + if (mysql_query ( h->conn, "SELECT u_token,t_space/" BYTES_IN_MEGABYTE_STR ", u_space/" BYTES_IN_MEGABYTE_STR " " + "FROM dpm_space_reserv " ) != 0) { + dpm_print_error( h->conn, "mysql_query() failed"); + return; + } + + results = mysql_store_result ( h->conn); /* generate result set */ + if (results == NULL) { + dpm_print_error ( h->conn, "mysql_store_result() failed"); + return; + } + + tokens_root_branch = datatree_new_branch ("tokens"); + datatree_add_node( root, tokens_root_branch); + + while( (row = mysql_fetch_row( results)) != NULL) { + + /* build new datatree leaf node */ +/* + result_tokens_total = datatree_new_result_from_string( "total", + UINT32_STR, + DATATREE_ATTR_NORMAL, + "MB", + row[1]); +*/ + +/* +void *datatree_new_result( const char *key, datatree_type_t type, + datatree_value_t value, + int attr, + const char *units); + +*/ + + total = strtoll(row[1], NULL, 10); + free = strtoll(row[2], NULL, 10); + used = total - free; + used_val.uintVal = used; + total_val.uintVal = total; + + result_tokens_used = datatree_new_result( "used", + UINT32, + used_val, + DATATREE_ATTR_NORMAL, + "MB"); + result_tokens_free = datatree_new_result_from_string( "free", + UINT32_STR, + DATATREE_ATTR_NORMAL, + "MB", + row[2]); + result_tokens_total = datatree_new_result( "total", + UINT32, + total_val, + DATATREE_ATTR_NORMAL, + "MB"); + + /* add leaf node to appropriate branch */ + datatree_add_node_with_path( tokens_root_branch, row[0], result_tokens_free, 0); + datatree_add_node_with_path( tokens_root_branch, row[0], result_tokens_used, 0); + datatree_add_node_with_path( tokens_root_branch, row[0], result_tokens_total, 0); + } + + + mysql_free_result(results); +} diff -ur monami.orig/MonAMI/src/dpm/getvalues.h monami/MonAMI/src/dpm/getvalues.h --- monami.orig/MonAMI/src/dpm/getvalues.h 2006-12-06 12:45:46.000000000 +0100 +++ monami/MonAMI/src/dpm/getvalues.h 2008-09-01 21:48:06.000000000 +0200 @@ -39,3 +39,4 @@ void dpm_get_usedvalues( dpm_handle_t *h, void *root); void dpm_get_overallvalues( dpm_handle_t *h, void *root); void dpm_get_fsspace( dpm_handle_t *h, void *root); +void dpm_get_tokens( dpm_handle_t *h, void *root); diff -ur monami.orig/MonAMI/src/dpm/interface.c monami/MonAMI/src/dpm/interface.c --- monami.orig/MonAMI/src/dpm/interface.c 2008-02-18 02:13:12.000000000 +0100 +++ monami/MonAMI/src/dpm/interface.c 2008-09-02 07:46:34.000000000 +0200 @@ -195,5 +195,6 @@ dpm_get_usedvalues(h, root); dpm_get_overallvalues(h, root); dpm_get_fsspace(h, root); + dpm_get_tokens(h, root); } |
From: Paul M. <p.m...@ph...> - 2008-09-12 08:25:38
|
Hi Tomas, On Thursday 11 September 2008 09:37:52 Tomas Kouba wrote: > I have done some patching for tokens support. The patch (against current > cvs) is attached. Looks good! I'm afraid I can't look at merging this until next week as I'm currently at Karlsruhe for the GridKa school. > It is a simple query to mysql database and patch for monami ganglia > externals. A few questions: It looks like you're reporting two metrics for space reservations: "used" and "free". Would it be useful to publish a "total" metric of the reservation? (I think the other plugins tend to err towards publishing too much information :-) The "used" metric is published as a 64-bit (signed) integer and "free" as unsigned 32-bit integer. Is this difference intentional? How do people feel about publishing the capacities as an unsigned 32-bit integer? If I've got my sums right, this would give a maximum value of 4PT. If that's a little too tight, we could report the capacity metrics in GB. (I'm a little reluctant to use 64-bit integers, if possible, because not all monitoring systems understands them; e.g., MonAlisa doesn't. There needs to be support for 64-bit numbers because MySQL publishes some metrics as 64-bit counters.) > You can see it in action here: > http://elf.farm.particle.cz/ganglia/mg-single-frame.php?c=Skurut_Core&h=dpm >1.egee.cesnet.cz&r=hour&fr=DPM > > (go to the bottom to see just one AUGERPROD token). Looks very nice! Cheers, Paul. |
From: Tomas K. <to...@ji...> - 2008-09-11 08:38:11
|
Tomas Kouba wrote: > 2) Has someone added support for space tokens into monami > dpm plugin? Hi, I have done some patching for tokens support. The patch (against current cvs) is attached. It is a simple query to mysql database and patch for monami ganglia externals. You can see it in action here: http://elf.farm.particle.cz/ganglia/mg-single-frame.php?c=Skurut_Core&h=dpm1.egee.cesnet.cz&r=hour&fr=DPM (go to the bottom to see just one AUGERPROD token). -- Tomas Kouba |
From: Paul M. <p.m...@ph...> - 2008-08-27 21:34:09
|
Hi Tomas, I've jotted some ideas below, as much to sound some ideas and to see what the others think (and also so people can correct the bits I've got wrong :-) On Wednesday 27 August 2008 18:17:26 Tomas Kouba wrote: > 1) Is there a plugin (or a person working on the plugin) > that would allow output to munin? > ( http://munin.projects.linpro.no/ ) Nope, currently there's no plugin for munin. But, having a quick look, adding a plugin looks to be possible. I don't know munin but skimming through the code it looks like fetching remote metrics by calling a Perl script called munin-gather: http://munin.projects.linpro.no/browser/trunk/server/munin-gather.in If so, then a remote agent (e.g., MonAMI) should listen on port TCP port 4949 for connections. If we ignoring TLS to begin with, the update process sends a command like: fetch <name> and return lines like (where "# comment" is optional) <metricname>.value <value> # comment <metricname>.value <date>:<value> # comment or, optionally, lines like: <metricname>.extinfo <location> This is an example of on-demand monitoring. It's slightly more involved than int.triggered monitoring (such as Ganglia) due to it's async. nature, but the ksysguard plugin does much the same so it shouldn't be too hard to knock this one together (hopefully!). Incidentally, I've got some skeleton code for supporting "abstracted" on-demand monitoring, (it needs some TLC before integrating into monami core). When it's working, it should provide the support needed for the sample, ksysguard and a new NRPE-like mode in the nagios plugin. From a MonAMI config point-of-view, I guess the simplest thing would be to have a munin plugin that accepts a list of read targets. For example: [torque] host = torque.example.org [maui] host = maui.example.org [munin] read = torque, maui allow = munin.example.org Then, the munin munin-gather could, after connecting on port 4949, send: fetch torque to receive all the torque metrics. Does this make sense? > 2) Has someone added support for space tokens into monami > dpm plugin? Sadly no. The existing plugin discovers information by querying the underlying MySQL database directly. This isn't the best as it locks the plugin into the DPM database schema. (a better approach is to query DPM via the DPM RFIO client-side library. I did some initial investigation here but, what with one thing and another, this ended up taking a back-seat.) I had a brief email exchange with Greig when he released his nice API / tool for querying DPM. I believe this, too, talks to the DPM databases. The code should include the queries necessary for pulling out token information. So, figuring out what SQL one should use is (more or less) solved. I've not had much spare time to work on the DPM plugin, but I'm very happy to help anyone who wants to add additional information. > I will try to implement these but don't want to duplicate someone's work > so I am asking here first. Very good! Please keep in touch with how you get on and say if you're not sure about anything. All the best, Paul. |
From: Andrew E. <and...@gm...> - 2008-08-27 17:07:54
|
> 1) Is there a plugin (or a person working on the plugin) > that would allow output to munin? ooooh - This would be nice as I use munin at home - I could then get currentcost to munin via monami :-) <aol>me too</aol> > 2) Has someone added support for space tokens into monami > dpm plugin? not AFAIK but +1 on the 'yes please' list Andrew |
From: Tomas K. <to...@ji...> - 2008-08-27 16:17:47
|
Hello all, I have two questions regarding future of monami: 1) Is there a plugin (or a person working on the plugin) that would allow output to munin? ( http://munin.projects.linpro.no/ ) 2) Has someone added support for space tokens into monami dpm plugin? I will try to implement these but don't want to duplicate someone's work so I am asking here first. Thank you, -- Tomas Kouba |
From: Peter L. <p....@la...> - 2008-07-08 09:11:14
|
Great, all fixed now thanks Tomas. In hindsight I was being stupid and didn't do the obvious thing as the error msg suggested. The fix is with ganglia-web by adding the include_once functions.php to get_context.php. Cheers, Peter Tomas Kouba (to...@ji...) wrote: > Peter Love wrote: > >Hey Phil, yeah you mentioned it but it's already been patched in > >multiple-graphs.php > > > >/** > > * The following include is to work-around broken deps in > > * Ganglia web front-end v3.0.7 (thanks are due to Phil Roffe) > > */ > >include_once "./functions.php"; > > > >Is there a need to add it to other php files? It doesn't seem to help. > > > >Peter > > Hi, > I have had a similar problem. I put debug_print_backtrace() before the > problematic clean_string() call. It showed me the source file that lacks the > include_once directive. > > -- > Tomas Kouba |
From: Tomas K. <to...@ji...> - 2008-07-08 07:10:17
|
Peter Love wrote: > Hey Phil, yeah you mentioned it but it's already been patched in > multiple-graphs.php > > /** > * The following include is to work-around broken deps in > * Ganglia web front-end v3.0.7 (thanks are due to Phil Roffe) > */ > include_once "./functions.php"; > > Is there a need to add it to other php files? It doesn't seem to help. > > Peter Hi, I have had a similar problem. I put debug_print_backtrace() before the problematic clean_string() call. It showed me the source file that lacks the include_once directive. -- Tomas Kouba |
From: Peter L. <p....@la...> - 2008-07-04 15:54:51
|
Hey Phil, yeah you mentioned it but it's already been patched in multiple-graphs.php /** * The following include is to work-around broken deps in * Ganglia web front-end v3.0.7 (thanks are due to Phil Roffe) */ include_once "./functions.php"; Is there a need to add it to other php files? It doesn't seem to help. Peter Phil Roffe (phi...@du...) wrote: > One more suggestion (which I think I mentioned before), I needed to add > the following at the top of the monami graph php files > (torque-graph.php, mysql-graph.php, dpm-graph.php, mg-frame-dpm.php etc). > > include_once "./functions.php"; > > Assuming the clean_string is in functions.php, can you try this? > > Cheers, > Phil > > --- > Phil Roffe - Phi...@du... > > IPPP, Department of Physics, Durham University, > Science Laboratories, South Road, Durham, DH1 3LE > Direct Dial: +44 (0)191 3343704 > Office: +44 (0)191 334 3811 > > > > Phil Roffe wrote: > > From a grep in the ganglia web directory it seems the function > > clean_string() is in functions.php > > > > [root@master ganglia]# grep "function clean_string" * > > functions.php:function clean_string( $string ) > > > > Can you try the same grep and check its there. If not, maybe you have a > > different ganglia-web version? > > > > [root@master ganglia]# rpm -qa | grep ganglia > > ganglia-gmond-3.0.7-1 > > ganglia-gmetad-3.0.7-1 > > ganglia-web-3.0.7-1 > > > > And version.php shows: > > $ganglia_version = "3.0.7"; > > > > Cheers, > > Phil > > > > Peter Love wrote: > > > >> Maybe someone can dig around for clean_string() then we can figure out > >> why our install can't find it. > >> > >> Paul Millar (p.m...@ph...) wrote: > >> > >> > >>> Hi Peter, > >>> > >>> On Tuesday 01 July 2008 18:20:07 Peter Love wrote: > >>> > >>> > >>>> Still haven't managed to crack this one, has anyone seen this problem in > >>>> thier httpd logs? > >>>> > >>>> PHP Fatal error: Call to undefined function: clean_string() in > >>>> /var/www/html/ganglia/get_contex > >>>> t.php on line 9 > >>>> > >>>> using: > >>>> ganglia 3.0.7 > >>>> php 4.3.9 > >>>> > >>>> > >>> Sorry, I didn't realise this was still a problem :-( > >>> > >>> I'm going to be testing some pending changes to the ganglia plugin. I'll have > >>> another go at trying to replicate the problem. > >>> > >>> Cheers, > >>> > >>> Paul. > >>> > >>> > > > > > > ------------------------------------------------------------------------- > > Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! > > Studies have shown that voting for your favorite open source project, > > along with a healthy diet, reduces your potential for chronic lameness > > and boredom. Vote Now at http://www.sourceforge.net/community/cca08 > > _______________________________________________ > > MonAMI-users mailing list > > Mon...@li... > > https://lists.sourceforge.net/lists/listinfo/monami-users > > |
From: Phil R. <phi...@du...> - 2008-07-04 08:08:07
|
One more suggestion (which I think I mentioned before), I needed to add the following at the top of the monami graph php files (torque-graph.php, mysql-graph.php, dpm-graph.php, mg-frame-dpm.php etc). include_once "./functions.php"; Assuming the clean_string is in functions.php, can you try this? Cheers, Phil --- Phil Roffe - Phi...@du... IPPP, Department of Physics, Durham University, Science Laboratories, South Road, Durham, DH1 3LE Direct Dial: +44 (0)191 3343704 Office: +44 (0)191 334 3811 Phil Roffe wrote: > From a grep in the ganglia web directory it seems the function > clean_string() is in functions.php > > [root@master ganglia]# grep "function clean_string" * > functions.php:function clean_string( $string ) > > Can you try the same grep and check its there. If not, maybe you have a > different ganglia-web version? > > [root@master ganglia]# rpm -qa | grep ganglia > ganglia-gmond-3.0.7-1 > ganglia-gmetad-3.0.7-1 > ganglia-web-3.0.7-1 > > And version.php shows: > $ganglia_version = "3.0.7"; > > Cheers, > Phil > > Peter Love wrote: > >> Maybe someone can dig around for clean_string() then we can figure out >> why our install can't find it. >> >> Paul Millar (p.m...@ph...) wrote: >> >> >>> Hi Peter, >>> >>> On Tuesday 01 July 2008 18:20:07 Peter Love wrote: >>> >>> >>>> Still haven't managed to crack this one, has anyone seen this problem in >>>> thier httpd logs? >>>> >>>> PHP Fatal error: Call to undefined function: clean_string() in >>>> /var/www/html/ganglia/get_contex >>>> t.php on line 9 >>>> >>>> using: >>>> ganglia 3.0.7 >>>> php 4.3.9 >>>> >>>> >>> Sorry, I didn't realise this was still a problem :-( >>> >>> I'm going to be testing some pending changes to the ganglia plugin. I'll have >>> another go at trying to replicate the problem. >>> >>> Cheers, >>> >>> Paul. >>> >>> > > > ------------------------------------------------------------------------- > Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! > Studies have shown that voting for your favorite open source project, > along with a healthy diet, reduces your potential for chronic lameness > and boredom. Vote Now at http://www.sourceforge.net/community/cca08 > _______________________________________________ > MonAMI-users mailing list > Mon...@li... > https://lists.sourceforge.net/lists/listinfo/monami-users > |
From: Phil R. <phi...@du...> - 2008-07-03 19:47:07
|
From a grep in the ganglia web directory it seems the function clean_string() is in functions.php [root@master ganglia]# grep "function clean_string" * functions.php:function clean_string( $string ) Can you try the same grep and check its there. If not, maybe you have a different ganglia-web version? [root@master ganglia]# rpm -qa | grep ganglia ganglia-gmond-3.0.7-1 ganglia-gmetad-3.0.7-1 ganglia-web-3.0.7-1 And version.php shows: $ganglia_version = "3.0.7"; Cheers, Phil Peter Love wrote: > Maybe someone can dig around for clean_string() then we can figure out > why our install can't find it. > > Paul Millar (p.m...@ph...) wrote: > >> Hi Peter, >> >> On Tuesday 01 July 2008 18:20:07 Peter Love wrote: >> >>> Still haven't managed to crack this one, has anyone seen this problem in >>> thier httpd logs? >>> >>> PHP Fatal error: Call to undefined function: clean_string() in >>> /var/www/html/ganglia/get_contex >>> t.php on line 9 >>> >>> using: >>> ganglia 3.0.7 >>> php 4.3.9 >>> >> Sorry, I didn't realise this was still a problem :-( >> >> I'm going to be testing some pending changes to the ganglia plugin. I'll have >> another go at trying to replicate the problem. >> >> Cheers, >> >> Paul. >> |
From: Peter L. <p....@la...> - 2008-07-03 10:01:30
|
Maybe someone can dig around for clean_string() then we can figure out why our install can't find it. Paul Millar (p.m...@ph...) wrote: > Hi Peter, > > On Tuesday 01 July 2008 18:20:07 Peter Love wrote: > > Still haven't managed to crack this one, has anyone seen this problem in > > thier httpd logs? > > > > PHP Fatal error: Call to undefined function: clean_string() in > > /var/www/html/ganglia/get_contex > > t.php on line 9 > > > > using: > > ganglia 3.0.7 > > php 4.3.9 > > Sorry, I didn't realise this was still a problem :-( > > I'm going to be testing some pending changes to the ganglia plugin. I'll have > another go at trying to replicate the problem. > > Cheers, > > Paul. |
From: Steve Y. <ch...@ha...> - 2008-07-03 00:57:06
|
Thanks Paul =). Exactly what I was looking for. -Steve On Jul 2, 2008, at 7:17 PM, Paul Millar wrote: > Hi Steve, > > On Tuesday 01 July 2008 21:17:53 Steve Young wrote: >> I was just wondering how does the efficiency of jobs get figured out >> from the torque output of "Running Jobs by Efficiency" as seen on >> this page: >> >> http://chem.hamilton.edu/modules/myiframe/index.php?iframeid=2 > > Yes, it's not immediately obvious, so I've added a section on this > within the > User Guide, ready for the next release. > > "Efficiency" is something the torque plugin calculates, rather than > something > the Torque server reports. It's a simple calculation: just divide > a job's CPU > time by it's wall-clock time (both metrics are reported by Torque > server). > This calculates (something like) the average efficiency of the job > over its > execution time so far. The number is then bracketed into five > "efficiency bins": > less than 20%, 20%--40%, 40%--60%, 60%--80% and greater than 80%. > What is > plotted is the number of running jobs in each bin. > > There's a couple of problems with this: one trivial the other > difficult. > > The first (easy) problem is that the above calculation doesn't take > into > account how many nodes a job is running on. The solution is to > count the > number of nodes a job is using and divide the efficiency by that > number. I've > fixed the code in CVS so the next version should give more correct > values. > > The second (hard) problem is that, if a job "busy-waits" for > something (most > likely network traffic) or is caught in a tight loop (i.e., a bug) > then the > process will consume lots of CPU, but not make any progress. So, > although it > appears to be 100% efficient, it might be making no progress > towards completing. > As it happens, MPI libraries tend to use busy-waits when waiting > for network > traffic, as it has lower overheads (provided there's no contention > for CPU > usage). > > So, the efficiency measurements should be taken with a pinch of > salt. If the > measured efficiency is low, then the job really is poorly utilising > the CPU and > making slow progress to completing its goal. If the measured > efficiency is high > then, unless you know the code isn't using busy-waits, you can't be > sure of > its efficiency. > > If the code does use busy-waits (like MPI jobs) then you need some > other means > of estimating efficiency. > >> Just thought I would start here... I'm guessing this might be >> something for the torque list. > > I've ask them ... which is where I got my information from ;-) > > Cheers, > > Paul. |
From: Paul M. <p.m...@ph...> - 2008-07-02 23:22:32
|
Hi Peter, On Tuesday 01 July 2008 18:20:07 Peter Love wrote: > Still haven't managed to crack this one, has anyone seen this problem in > thier httpd logs? > > PHP Fatal error: Call to undefined function: clean_string() in > /var/www/html/ganglia/get_contex > t.php on line 9 > > using: > ganglia 3.0.7 > php 4.3.9 Sorry, I didn't realise this was still a problem :-( I'm going to be testing some pending changes to the ganglia plugin. I'll have another go at trying to replicate the problem. Cheers, Paul. |
From: Paul M. <p.m...@ph...> - 2008-07-02 23:16:49
|
Hi Steve, On Tuesday 01 July 2008 21:17:53 Steve Young wrote: > I was just wondering how does the efficiency of jobs get figured out > from the torque output of "Running Jobs by Efficiency" as seen on > this page: > > http://chem.hamilton.edu/modules/myiframe/index.php?iframeid=2 Yes, it's not immediately obvious, so I've added a section on this within the User Guide, ready for the next release. "Efficiency" is something the torque plugin calculates, rather than something the Torque server reports. It's a simple calculation: just divide a job's CPU time by it's wall-clock time (both metrics are reported by Torque server). This calculates (something like) the average efficiency of the job over its execution time so far. The number is then bracketed into five "efficiency bins": less than 20%, 20%--40%, 40%--60%, 60%--80% and greater than 80%. What is plotted is the number of running jobs in each bin. There's a couple of problems with this: one trivial the other difficult. The first (easy) problem is that the above calculation doesn't take into account how many nodes a job is running on. The solution is to count the number of nodes a job is using and divide the efficiency by that number. I've fixed the code in CVS so the next version should give more correct values. The second (hard) problem is that, if a job "busy-waits" for something (most likely network traffic) or is caught in a tight loop (i.e., a bug) then the process will consume lots of CPU, but not make any progress. So, although it appears to be 100% efficient, it might be making no progress towards completing. As it happens, MPI libraries tend to use busy-waits when waiting for network traffic, as it has lower overheads (provided there's no contention for CPU usage). So, the efficiency measurements should be taken with a pinch of salt. If the measured efficiency is low, then the job really is poorly utilising the CPU and making slow progress to completing its goal. If the measured efficiency is high then, unless you know the code isn't using busy-waits, you can't be sure of its efficiency. If the code does use busy-waits (like MPI jobs) then you need some other means of estimating efficiency. > Just thought I would start here... I'm guessing this might be > something for the torque list. I've ask them ... which is where I got my information from ;-) Cheers, Paul. |
From: Steve Y. <ch...@ha...> - 2008-07-01 19:18:09
|
Hi, I was just wondering how does the efficiency of jobs get figured out from the torque output of "Running Jobs by Efficiency" as seen on this page: http://chem.hamilton.edu/modules/myiframe/index.php?iframeid=2 Just thought I would start here... I'm guessing this might be something for the torque list. -Steve |
From: Peter L. <p....@la...> - 2008-07-01 16:20:17
|
Still haven't managed to crack this one, has anyone seen this problem in thier httpd logs? PHP Fatal error: Call to undefined function: clean_string() in /var/www/html/ganglia/get_contex t.php on line 9 using: ganglia 3.0.7 php 4.3.9 Cheers, Peter Phil Roffe (phi...@du...) wrote: > Hi Peter, > > When I first installed the torque graphs were missing too, but I had the > following error... > *Fatal error*: Call to undefined function: clean_string() in > */export/docroot/www/ganglia/get_context.php* on line *9 > > *I fixed it by adding the following line to the top of torque-graph.php... > > include_once "./functions.php"; > > > I can't get the error msg on from your web... maybe a php setting? Can > you try adding the include line above and see if it fixes it? > > Thanks, > Phil > > --- > Phil Roffe - Phi...@du... > > IPPP, Department of Physics, Durham University, > Science Laboratories, South Road, Durham, DH1 3LE > Direct Dial: +44 (0)191 3343704 > Office: +44 (0)191 334 3811 > > > > Peter Love wrote: > >No luck with graph content: http://tinyurl.com/6xamq4 > > > >multiple-graphs.php function cmd_go( $cmd, $debug=1) > > > >doesn't spit out the usual commandline, any idea hoe to proceed? > > > >Paul Millar (p.m...@ph...) wrote: > > > >>Hi Peter, > >> > >>On Friday 23 May 2008 16:09:54 Peter Love wrote: > >> > >>>Paul, are the php script in any rpm packages? In particular, torque and > >>>dpm. We've setup a new MON box. Previously I got these from CVS head. > >>> > >>Sorry, they're currently not packaged; the only way to get them is from > >>CVS. > >>I was reluctant to package them as I wasn't sure into which directly the > >>files should go (although /var/www/ganglia seems a pretty safe bet) and > >>because installing them requires manual patching of Ganglia's web pages > >>(something I think it would be unwise to do automatically, though > >>post-install script). > >> > >>On a related note, I've recently updated the patches, based on feedback > >>from Phil Roffe (thanks Phil!). The CVS head now includes support for > >>Ganglia v3.0.7 web pages (by including an alternative patch). This > >>should work, but Phil reported that there were still some issues from PHP > >>not including the necessary files. > >> > >>HTH, > >> > >>Paul. > >> > > > >------------------------------------------------------------------------- > >This SF.net email is sponsored by: Microsoft > >Defy all challenges. Microsoft(R) Visual Studio 2008. > >http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > >_______________________________________________ > >MonAMI-users mailing list > >Mon...@li... > >https://lists.sourceforge.net/lists/listinfo/monami-users > > |