Menu
▾
▴
ganglia-bugzilla — Mailing list for receiving bugzilla submissions
You can subscribe to this list here.
| 2006 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(76) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2007 |
Jan
(32) |
Feb
(32) |
Mar
(58) |
Apr
(49) |
May
(34) |
Jun
(36) |
Jul
(46) |
Aug
(36) |
Sep
(32) |
Oct
(56) |
Nov
(64) |
Dec
(40) |
| 2008 |
Jan
(51) |
Feb
(58) |
Mar
(54) |
Apr
(38) |
May
(41) |
Jun
(47) |
Jul
(60) |
Aug
(70) |
Sep
(41) |
Oct
(54) |
Nov
(42) |
Dec
(32) |
| 2009 |
Jan
(55) |
Feb
(28) |
Mar
(34) |
Apr
(42) |
May
(25) |
Jun
(37) |
Jul
(37) |
Aug
(36) |
Sep
(32) |
Oct
(33) |
Nov
(24) |
Dec
|
| 2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(23) |
Jul
(11) |
Aug
(9) |
Sep
(24) |
Oct
(35) |
Nov
(18) |
Dec
(3) |
| 2011 |
Jan
(8) |
Feb
(16) |
Mar
(18) |
Apr
(13) |
May
(4) |
Jun
(2) |
Jul
(1) |
Aug
(6) |
Sep
(5) |
Oct
(5) |
Nov
(1) |
Dec
(6) |
| 2012 |
Jan
(1) |
Feb
(8) |
Mar
(8) |
Apr
(7) |
May
(6) |
Jun
(1) |
Jul
(2) |
Aug
|
Sep
(2) |
Oct
(4) |
Nov
(2) |
Dec
(1) |
| 2013 |
Jan
(2) |
Feb
(9) |
Mar
(5) |
Apr
(1) |
May
(5) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
| 2014 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2016 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(1) |
Oct
|
Nov
|
Dec
|
| 2017 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2013-01-15 17:44:59
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=341 Summary: diskstat.py does not recognise HP Smart Array P410i controller Product: Ganglia Monitoring System Version: 3.1.x Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P5 Component: gmond AssignedTo: gan...@li... ReportedBy: s.s...@sc... Estimated Hours: 0.0 The python module diskstat.py from gmond 3.3.7 does not recognise disks connected to a HP Smart Array P410i controller as in HP ProLiant DL380 servers. Operating system is Suse linux enterprise server 10. cat /proc/partitions major minor #blocks name 104 0 2050550062 cciss/c0d0 104 1 72261 cciss/c0d0p1 ... The first partition's device file is: /sys/block/cciss!c0d0/cciss!c0d0p1/ (with the exclamation marks) There are two problems with the python module: 1. The drive is ignored due to the following awk expression in diskstat.py: awk_cmd = "awk 'NR > 1 && $0 !~ /" + IGNORE_DEV + "/ && $4 !~ /[0-9]$/ {ORS=\" \"; print $4}' " (the critical part is "$4 !~ /[0-9]$/") 2. The path /sys/block/cciss!c0d0/cciss!c0d0p1/ is not found by the open command in "f = open('/sys/block/' + dev + '/size', 'r')" This completely uncritical. I can work around it in my special case. Just wanted to let you know. Best regards Stephan -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-12-29 14:08:22
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=340 Summary: Undefined symbol "cpu_steal_func" Product: Ganglia Monitoring System Version: trunk Platform: PC OS/Version: FreeBSD Status: NEW Severity: critical Priority: P3 Component: gmond AssignedTo: gan...@li... ReportedBy: as...@ev... Estimated Hours: 0.0 FreeBSD freebsd-doc 9.0-RELEASE FreeBSD 9.0-RELEASE #0: Tue Jan 3 07:15:25 UTC 2012 ro...@ob...:/usr/obj/usr/src/sys/GENERIC i386 ganglia-3.5.0 ./gmond Cannot load /opt/ganglia/lib/ganglia/modcpu.so metric module: /opt/ganglia/lib/ganglia/modcpu.so: Undefined symbol "cpu_steal_func" # nm modcpu.so | grep " cpu.*" 00002040 T cpu_aidle_func 000039e0 T cpu_idle_func 00003990 T cpu_intr_func 000017c0 t cpu_metric_cleanup 000017d0 t cpu_metric_handler 00005960 d cpu_metric_info 00001920 t cpu_metric_init 00005920 D cpu_module 00003a80 T cpu_nice_func 000024b0 T cpu_num_func 00002060 T cpu_sintr_func 00003f00 T cpu_speed_func 000037a0 T cpu_state 00005bc0 b cpu_states.6678 U cpu_steal_func 00003a30 T cpu_system_func 00003ad0 T cpu_user_func 00002020 T cpu_wio_func -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-11-29 17:57:08
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=339 Summary: web 2.0 can't open rrd file if node name has upper case chars Product: Ganglia Monitoring System Version: web 2.0 Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P5 Component: web frontend AssignedTo: gan...@li... ReportedBy: as...@ev... Estimated Hours: 0.0 ERROR: opening '/opt/ganglia/rrds/Clustername/NODE-Name/load_one.rrd': No such file or directory -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-11-29 17:57:03
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=339 --- Comment #1 from yuriy.tereschuk <as...@ev...> 2012-11-29 09:44:49 --- solution was found in creation symbolic link to destination folder ln -s /opt/ganglia/rrds/Clustername/node-name /opt/ganglia/rrds/Clustername/NODE-Name Looks like gmetad creates destination folders only with chars in lowercase. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-10-11 13:53:23
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=338 Summary: Add a nagios status button to host and node view Product: Ganglia Monitoring System Version: web 2.0 Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P5 Component: web frontend AssignedTo: gan...@li... ReportedBy: ke...@um... Estimated Hours: 0.0 Created an attachment (id=282) --> (http://bugzilla.ganglia.info/cgi-bin/bugzilla/attachment.cgi?id=282) Patch to add a nagios status button to host and node view The enclosed patch adds a 'Nagios Status' button to the host and node views which will take you directly to the nagios status page for that host. This also includes two configuration parameters: $conf['nagios_status_button'] a boolean to enable the button, which defaults to false $conf['nagios_status_url'] the URL for your nagios status CGI -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-10-11 13:27:49
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=337 Summary: Show reports only if their associated metrics are defined Product: Ganglia Monitoring System Version: web 2.0 Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P5 Component: web frontend AssignedTo: gan...@li... ReportedBy: ke...@um... Estimated Hours: 0.0 Created an attachment (id=281) --> (http://bugzilla.ganglia.info/cgi-bin/bugzilla/attachment.cgi?id=281) Patch to show reports only if their associated metrics are defined We have a large number of custom reports but only some of our monitored hosts collect metrics appropriate for those reports. The current web front end shows an empty graph when the metric isn't present. I'm enclosing a patch which adds a configuration array parameter ($report_to_metric) to associate metrics with reports. In host view, if the specified metric isn't defined for a given host, then the report won't be shown. In cluster view, if the specified metric isn't defined for the first host in the cluster, then the cluster summary report won't be shown. (Ideally this should probably be handled by the report module itself, but since the report module isn't called until after the page layout has already been determined, this isn't currently feasible) -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-10-11 12:32:51
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=336 Kevin Hildebrand <ke...@um...> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #279 is|0 |1 obsolete| | --- Comment #1 from Kevin Hildebrand <ke...@um...> 2012-10-11 05:32:39 --- Created an attachment (id=280) --> (http://bugzilla.ganglia.info/cgi-bin/bugzilla/attachment.cgi?id=280) Corrected patch to move meta_designator to $conf -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-10-11 12:27:52
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=336 Summary: Allow sites to configure their own meta designator Product: Ganglia Monitoring System Version: web 2.0 Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P5 Component: web frontend AssignedTo: gan...@li... ReportedBy: ke...@um... Estimated Hours: 0.0 Created an attachment (id=279) --> (http://bugzilla.ganglia.info/cgi-bin/bugzilla/attachment.cgi?id=279) Patch to move meta_designator to $conf We are using ganglia to monitor hundreds of machines across our campus, and would prefer to have the meta level be called 'Campus' instead of 'Grid'. I've attached patches that will take the previous global $meta_designator and move it into $conf[]. These patches also fix the one place where 'Grid' was hard coded instead of using $meta_designator. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-09-26 07:40:21
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=317 Clotho Tsang <clo...@gm...> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |clo...@gm... -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-09-19 13:42:43
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=335 Summary: Add graph to view error Product: Ganglia Monitoring System Version: web 2.0 Platform: All OS/Version: Linux Status: NEW Severity: major Priority: P4 Component: web frontend AssignedTo: gan...@li... ReportedBy: oli...@sy... Estimated Hours: 0.0 When clicking on a graph to add this graph to a view, after selecting an exisitng view, an error display : Alert : view does not exist. This should not happen. In the view menu, I can see the view name. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-07-15 21:44:09
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=334 Kostas Georgiou <k.g...@at...> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |k.g...@at... --- Comment #1 from Kostas Georgiou <k.g...@at...> 2012-07-15 14:44:00 --- In the fedora/epel rpms it is a link (private_clusters -> ../../../etc/ganglia/private_clusters) so httpd throws a "Symbolic link not allowed or link target not accessible: /usr/share/ganglia/private_clusters" if someone tries to access it from the web. Maybe something similar could be done to stop it from being accessible by everyone. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-07-15 21:26:01
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=333 Bernard Li <be...@va...> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |be...@va... AssignedTo|ganglia- |be...@va... |bug...@li...urceforge.n| |et | -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-06-21 20:54:50
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=317 gernot schilling <ger...@gm...> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ger...@gm... --- Comment #1 from gernot schilling <ger...@gm...> 2012-06-21 13:54:42 --- this seems to be similar to the bugs http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=238 http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=251 i had the same problem, this patch fixed it for me: (the other bugs talked about uint64_t, but as total_size and total_free are already doubles, i used that) --- ganglia-3.4.0/libmetrics/linux/metrics.c 2012-05-02 22:20:36.000000000 +0200 +++ ganglia-3.4.0-fs/libmetrics/linux/metrics.c 2012-06-21 22:44:49.432917578 +0200 @@ -1174,9 +1174,9 @@ float device_space(char *mount, char *device, double *total_size, double *total_free) { struct statvfs svfs; - uint32_t blocksize; - uint32_t free; - uint32_t size; + double blocksize; + double free; + double size; /* The percent used: used/total * 100 */ float pct=0.0; @@ -1192,8 +1192,8 @@ size = svfs.f_blocks; blocksize = svfs.f_bsize; /* Keep running sum of total used, free local disk space. */ - *total_size += size * (double) blocksize; - *total_free += free * (double) blocksize; + *total_size += size * blocksize; + *total_free += free * blocksize; /* The percentage of space used on this partition. */ pct = size ? ((size - free) / (float) size) * 100 : 0.0; return pct; -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-05-31 11:49:43
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=334 Summary: Password Weaknesses Product: Ganglia Monitoring System Version: 3.1.x Platform: PC OS/Version: All Status: NEW Severity: minor Priority: P5 Component: web frontend AssignedTo: gan...@li... ReportedBy: ar...@em... Estimated Hours: 0.0 The following password weaknesses were discovered in version 3.1.7. Subsequent versions may have fixed these problems, but they are provided here so that it can be verified that all of them have been fixed. 1) By default, the private_clusters file could be accessed without being authenticated. This file lists all the private clusters and their corresponding password hashes. 2) Passwords for private clusters and the control room were sent using basic authentication (only Base64 encoding). Therefore, the passwords are essentially in clear text and can be easily obtained by sniffing network traffic. 3) Several factors made the private cluster passwords more susceptible to password cracking: 3a) The user name was ignored when trying to access the control room or private clusters. 3b)Users do not seem to be locked out for too many failed login attempts to a private cluster or to the control room. 3c) Password hashes can be obtained from the private_clusters file without being authenticated. These hashes are standard MD5 hashes; no salting is used. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-05-31 11:32:00
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=333 Summary: Directory Traversal Possible in graph.php Product: Ganglia Monitoring System Version: 3.1.x Platform: PC OS/Version: All Status: NEW Severity: major Priority: P2 Component: web frontend AssignedTo: gan...@li... ReportedBy: ar...@em... Estimated Hours: 0.0 Arbitrary PHP files may be executed from the Ganglia server due to unsanitized input to graph.php. Browsing to http://ganglia_server/graph.php?g=../../../../../$SOMEPATH/test results in the execution of $SOMEPATH/test.php (if the file exists and both it and its containing directories are world readable). $SOMEPATH does not need to be in the WEB_ROOT directory. After the code of the target PHP file is rendered, an error is often displayed, assuming the PHP file was not a valid graph. This vulnerability was discovered in version 3.1.7, but a cursory look at later versions suggests that it is still present. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-05-31 00:40:54
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=332 Summary: Segfault in gmetad with long "data source" line Product: Ganglia Monitoring System Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P5 Component: gmetad AssignedTo: gan...@li... ReportedBy: qha...@gm... Estimated Hours: 0.0 I have found with gmetad 3.4.0 if the "data source" line in the config file is sufficiently long it will cause a segfault in gmetad immediately after reading the file. I haven't found the lower limit of what causes the segfault, but a cluster of 40 nodes with hostnames of about 50 characters each with ports defined is sufficiently long to make it happen. That works out to be about 2160 characters, not counting spaces. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-05-30 15:50:41
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=258 Thomas Robitaille <tho...@gm...> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tho...@gm... --- Comment #2 from Thomas Robitaille <tho...@gm...> 2012-05-30 08:50:29 --- I have the same issue with MacOS 10.7, and applying the changes in the patch works for me (I couldn't apply the patch directly for some reason). -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-05-24 13:06:56
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=324 Aaron Nichols <ani...@tr...> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ani...@tr... --- Comment #3 from Aaron Nichols <ani...@tr...> 2012-05-24 06:06:47 --- I'll just add that I too have experienced this problem - I just haven't had cycles to try to get someones attention. There was also a pull request submitted to fix this: https://github.com/ganglia/monitor-core/pull/35 Likely related, when I have gmetad's configured to poll other gmetads (they do poll, the results just never get written to rrds) the polling gmetad instance will segfault once or twice a day in our environment. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-05-03 19:53:27
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=331 Summary: redis python module causes gmond to hang - 3.3.6 Product: Ganglia Monitoring System Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P5 Component: gmond AssignedTo: gan...@li... ReportedBy: chr...@gm... Estimated Hours: 0.0 on line 10 of redis.py s.send("INFO\n") Redis expects "\r\n" So gmond will hang on the following line s.recv(4096) change to: s.send("INFO\r\n") and all is good. This was tested on CentOS 5.8 and redis 2.4.11 with Ganglia 3.3.6 -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-04-24 20:44:39
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=330 Summary: gmond 3.3.5 segfaults Product: Ganglia Monitoring System Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P5 Component: gmond AssignedTo: gan...@li... ReportedBy: dh...@al... Estimated Hours: 0.0 Created an attachment (id=278) --> (http://bugzilla.ganglia.info/cgi-bin/bugzilla/attachment.cgi?id=278) gmond.conf + gmond debug output A simple multicast config (1 udp_send, 1 udp_recv and 1 tcp_accept) results in segfault. gmond runs for long enough to report some stats and for a gmetad to query it, but segfaults within a couple of minutes or less. Linux cdn.zlvpgh.com 2.6.32-220.2.1.el6.centos.plus.x86_64 #1 SMP Thu Dec 22 23:32:31 GMT 2011 x86_64 x86_64 x86_64 GNU/Linux gmond 3.3.5 -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-04-20 21:47:06
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=321 Daniel Pocock <da...@po...> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Daniel Pocock <da...@po...> 2012-04-20 14:46:57 --- Added a check for NULL in the 3.3.7 release -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-04-18 05:26:42
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=329 Summary: gmond memory leaks when deleting stale gmetric data Product: Ganglia Monitoring System Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P5 Component: gmond AssignedTo: gan...@li... ReportedBy: tan...@ho... Estimated Hours: 0.0 In gmond.c, function Ganglia_metadata_save(...) and Ganglia_value_save(...): metric and metric->name memory is allocated from host->pool. When gmetric data timed out and is deleted in Cleanup_data(...), metric is removed from gmetric hash table and metric->pool is destroyed. But allocated memory for metric and metric->name can't be released unless host->pool is destroyed. If gmond receive same gmetric data again, metric and metric->name alway allocate new memory. If certain gmetric data switchs from time out/active frequently, memory usage can increase more than 40M per day. My temporary solution is to comment out following code in Cleanup_data to avoid allocating memory again: //apr_hash_set( host->metrics, metric->name, APR_HASH_KEY_STRING, NULL); //apr_hash_set( host->gmetrics, metric->name, APR_HASH_KEY_STRING, NULL); /* destroy any memory that was allocated for this gmetric */ //apr_pool_destroy( metric->pool ); I think metric->name can be allocated from metric->pool, but metric may need a new pool to hold. Thanks Wei -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-04-10 17:11:08
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=328 --- Comment #2 from David Kowis <dk...@sh...> 2012-04-10 10:10:56 --- Okay, so I figured out a solution for me, probably specific to Source Mage GNU/Linux. Solution was to sed -i "s/apr-1/apr/g" configure before running ./configure Everything works fantastically after that. It'd be nice if the build system was capable of knowing about the proper way to link to apr based on the output of apr-1-config, but this is good enough for me. I'll leave it here and leave it up to you guys to close if you want. Thanks anyway! -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-04-10 17:01:20
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=328 --- Comment #1 from David Kowis <dk...@sh...> 2012-04-10 10:01:06 --- Tried running ./bootstrap to get it to regenerate all the everything, even that didn't work :( -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |
|
From: bugzilla-daemon@Millennium.Berkeley.EDU - 2012-04-10 16:49:30
|
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=328 Summary: Cannot build when apr 1.x isn't linked using -lapr-1 Product: Ganglia Monitoring System Version: 3.1.x Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P5 Component: build AssignedTo: gan...@li... ReportedBy: dk...@sh... Estimated Hours: 0.0 On Source Mage linux, apr 1.x is linked using -lapr not -lapr-1. Somehow the autoconfiguration is always coming up with -lapr-1, as well as the default ./configure script. I've asked around on #ganglia in irc trying to figure out what the necessary magic is to get it to properly build. I have tried passing in ./configure --with-libapr=/usr/bin/apr-1-config to specify the path to apr-1-config and try to force it to use the proper lib output. apr-1-config puts out the right information: apr-1-config --link-ld -L/usr/lib -lapr But somehow it's not being used or respected, and so I'm having trouble building the thing. I've tried using sed to replace all -lapr-1 with -lapr, and that works, until make is run again, at which point it runs aclocal and that completely clobbers it again :( Thanks for any help. -- Configure bugmail: http://bugzilla.ganglia.info/cgi-bin/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. |