Re: [Gluster-devel] glustershd status
On Wed, Jul 16, 2014 at 09:54:49PM -0700, Harshavardhana wrote: In-fact this is true on Linux as well - there is smaller time window observe the below output , immediately run 'volume status' after a 'volume start' event I observe the same lapse on NetBSD if the volume is created and started. If it is stoped and started again, glustershd will never report back. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] glustershd status
On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! On Wed, Jul 16, 2014 at 11:29 PM, Emmanuel Dreyfus m...@netbsd.org wrote: On Wed, Jul 16, 2014 at 09:54:49PM -0700, Harshavardhana wrote: In-fact this is true on Linux as well - there is smaller time window observe the below output , immediately run 'volume status' after a 'volume start' event I observe the same lapse on NetBSD if the volume is created and started. If it is stoped and started again, glustershd will never report back. -- Emmanuel Dreyfus m...@netbsd.org -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] glustershd status
Emmanuel, Could you take statedump* of the glustershd process when it has leaked enough memory to be able to observe and share the output? This might give us what kind of objects are we allocating abnormally high. * statedump of a glusterfs process #kill -USR1 pid of process HTH, Krish - Original Message - On Wed, Jul 16, 2014 at 11:32:06PM -0700, Harshavardhana wrote: On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! I observed memory leaks on long terme usage. My favourite test case is building NetBSD on a replicated/distributed volume, and I can see processes growing a lot during the build. I reported it some time ago, and some leaks were plugged, but obviosuly some remain. valgrind was never ported to NetBSD, hence I lack investigative tools, but I bet the leaks exist on FreeBSD and Linux as well. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] glustershd status
KP, I do have a 3.2Gigs worth of valgrind output which indicates this issue, trying to reproduce this on Linux. My hunch says that 'compiling' with --disable-epoll might actually trigger this issue on Linux too. Will update here once i have done that testing. On Wed, Jul 16, 2014 at 11:44 PM, Krishnan Parthasarathi kpart...@redhat.com wrote: Emmanuel, Could you take statedump* of the glustershd process when it has leaked enough memory to be able to observe and share the output? This might give us what kind of objects are we allocating abnormally high. * statedump of a glusterfs process #kill -USR1 pid of process HTH, Krish - Original Message - On Wed, Jul 16, 2014 at 11:32:06PM -0700, Harshavardhana wrote: On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! I observed memory leaks on long terme usage. My favourite test case is building NetBSD on a replicated/distributed volume, and I can see processes growing a lot during the build. I reported it some time ago, and some leaks were plugged, but obviosuly some remain. valgrind was never ported to NetBSD, hence I lack investigative tools, but I bet the leaks exist on FreeBSD and Linux as well. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] glustershd status
Harsha, In addition to the valgrind output, statedump output of glustershd process when the leak is observed would be really helpful. thanks, Krish - Original Message - Nope spoke too early, using poll() has no effect on the memory usage on Linux, so actually back to FreeBSD. On Thu, Jul 17, 2014 at 12:07 AM, Harshavardhana har...@harshavardhana.net wrote: KP, I do have a 3.2Gigs worth of valgrind output which indicates this issue, trying to reproduce this on Linux. My hunch says that 'compiling' with --disable-epoll might actually trigger this issue on Linux too. Will update here once i have done that testing. On Wed, Jul 16, 2014 at 11:44 PM, Krishnan Parthasarathi kpart...@redhat.com wrote: Emmanuel, Could you take statedump* of the glustershd process when it has leaked enough memory to be able to observe and share the output? This might give us what kind of objects are we allocating abnormally high. * statedump of a glusterfs process #kill -USR1 pid of process HTH, Krish - Original Message - On Wed, Jul 16, 2014 at 11:32:06PM -0700, Harshavardhana wrote: On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! I observed memory leaks on long terme usage. My favourite test case is building NetBSD on a replicated/distributed volume, and I can see processes growing a lot during the build. I reported it some time ago, and some leaks were plugged, but obviosuly some remain. valgrind was never ported to NetBSD, hence I lack investigative tools, but I bet the leaks exist on FreeBSD and Linux as well. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Developer Documentation for datastructures in gluster
On 07/15/2014 04:39 PM, Pranith Kumar Karampuri wrote: hi, Please respond if you guys volunteer to add documentation for any of the following things that are not already taken. client_t - pranith integration with statedump - pranith mempool - Pranith event-hostory + circ-buff - Raghavendra Bhat inode - Raghavendra Bhat call-stub fd iobuf graph xlator option-framework rbthash runner-framework stack/frame strfd timer store gid-cache(source is heavily documented) dict event-poll I'll take up event-poll. I have created an etherpad link with the components and volunteers thus far: https://etherpad.wikimedia.org/p/glusterdoc Feel free to update this doc with your patch details, other components etc. - Ravi Pranith ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
[Gluster-devel] Unnecessary bugs entered in BZ, ...
...and I'm sorry about that, the following documents are somewhat contradictory: http://gluster.org/community/documentation/index.php/Simplified_dev_workflow The script will ask you to enter a bugzilla bug id. Every change submitted to GlusterFS needs a bugzilla entry to be accepted. If you do not already have a bug id, file a new bug at Red Hat Bugzilla. If the patch is submitted for review, the rfc.sh script will return the gerrit url for the review request. www.gluster.org/community/documentation/index.php/Development_Work_Flow Prompt for a Bug Id for each commit (if it was not already provded) and include it as a BUG: tag in the commit log. You can just hit enter at this prompt if your submission is purely for review purposes. /Anders -- Anders Blomdell Email: anders.blomd...@control.lth.se Department of Automatic Control Lund University Phone:+46 46 222 4625 P.O. Box 118 Fax: +46 46 138118 SE-221 00 Lund, Sweden ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Unnecessary bugs entered in BZ, ...
I don't understand your confusion, but maybe these docs need to be reworded. What both of these want to say is that for a patch to be merged into glusterfs, it needs to be associated with a bug-id. This association is done by adding a 'BUG: id' line in the commit message. If you haven't manually added a bug-id in the commit message, the rfc.sh script will prompt you to enter one and add it to the commit-message. But, it is possible ignore this prompt and submit a patch for review. A patch submitted for review in this manner will only be reviewed. It will not be merged. The simplified workflow document doesn't mention this as it was targeted at new developers, and I felt having this details was TMI for them. But now when I rethink it, it's the seasoned developers who are beginning to contribute to gluster, who are more likely to use that doc. ~kaushal On Thu, Jul 17, 2014 at 6:19 PM, Anders Blomdell anders.blomd...@control.lth.se wrote: ...and I'm sorry about that, the following documents are somewhat contradictory: http://gluster.org/community/documentation/index.php/Simplified_dev_workflow The script will ask you to enter a bugzilla bug id. Every change submitted to GlusterFS needs a bugzilla entry to be accepted. If you do not already have a bug id, file a new bug at Red Hat Bugzilla. If the patch is submitted for review, the rfc.sh script will return the gerrit url for the review request. www.gluster.org/community/documentation/index.php/Development_Work_Flow Prompt for a Bug Id for each commit (if it was not already provded) and include it as a BUG: tag in the commit log. You can just hit enter at this prompt if your submission is purely for review purposes. /Anders -- Anders Blomdell Email: anders.blomd...@control.lth.se Department of Automatic Control Lund University Phone:+46 46 222 4625 P.O. Box 118 Fax: +46 46 138118 SE-221 00 Lund, Sweden ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Unnecessary bugs entered in BZ, ...
On 2014-07-17 15:44, Kaushal M wrote: I don't understand your confusion, but maybe these docs need to be reworded. Not confused, just observed that some of my patches regarding development workflow does not need to show up as bugs for gluster (at least not until somebody has verified that they could be useful to somebody else but me). What both of these want to say is that for a patch to be merged into glusterfs, it needs to be associated with a bug-id. This association is done by adding a 'BUG: id' line in the commit message. If you haven't manually added a bug-id in the commit message, the rfc.sh script will prompt you to enter one and add it to the commit-message. But, it is possible ignore this prompt and submit a patch for review. A patch submitted for review in this manner will only be reviewed. It will not be merged. The simplified workflow document doesn't mention this as it was targeted at new developers, and I felt having this details was TMI for them. But now when I rethink it, it's the seasoned developers who are beginning to contribute to gluster, who are more likely to use that s/seasoned developers/ignorant old fools/g :-) doc. ~kaushal On Thu, Jul 17, 2014 at 6:19 PM, Anders Blomdell anders.blomd...@control.lth.se wrote: ...and I'm sorry about that, the following documents are somewhat contradictory: http://gluster.org/community/documentation/index.php/Simplified_dev_workflow The script will ask you to enter a bugzilla bug id. Every change submitted to GlusterFS needs a bugzilla entry to be accepted. If you do not already have a bug id, file a new bug at Red Hat Bugzilla. If the patch is submitted for review, the rfc.sh script will return the gerrit url for the review request. www.gluster.org/community/documentation/index.php/Development_Work_Flow Prompt for a Bug Id for each commit (if it was not already provded) and include it as a BUG: tag in the commit log. You can just hit enter at this prompt if your submission is purely for review purposes. /Anders -- Anders Blomdell Email: anders.blomd...@control.lth.se Department of Automatic Control Lund University Phone:+46 46 222 4625 P.O. Box 118 Fax: +46 46 138118 SE-221 00 Lund, Sweden ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Unnecessary bugs entered in BZ, ...
Bugs/bug in the documents are not really what you think they mean. We use bugzilla to track glusterfs development. It is used for track defects (bugs) in the software, enhancements and new features. What we really mean by a patch needs a bug id is that we need it to have an entry on bugzilla, so that it's lifecycle can be tracked correctly. So bug just means bugzilla entry. You have bug bugs, enhancement bugs, feature request bugs and so on. tl;dr: bug == bugzilla entry. On Thu, Jul 17, 2014 at 8:58 PM, Anders Blomdell anders.blomd...@control.lth.se wrote: On 2014-07-17 15:44, Kaushal M wrote: I don't understand your confusion, but maybe these docs need to be reworded. Not confused, just observed that some of my patches regarding development workflow does not need to show up as bugs for gluster (at least not until somebody has verified that they could be useful to somebody else but me). What both of these want to say is that for a patch to be merged into glusterfs, it needs to be associated with a bug-id. This association is done by adding a 'BUG: id' line in the commit message. If you haven't manually added a bug-id in the commit message, the rfc.sh script will prompt you to enter one and add it to the commit-message. But, it is possible ignore this prompt and submit a patch for review. A patch submitted for review in this manner will only be reviewed. It will not be merged. The simplified workflow document doesn't mention this as it was targeted at new developers, and I felt having this details was TMI for them. But now when I rethink it, it's the seasoned developers who are beginning to contribute to gluster, who are more likely to use that s/seasoned developers/ignorant old fools/g :-) doc. ~kaushal On Thu, Jul 17, 2014 at 6:19 PM, Anders Blomdell anders.blomd...@control.lth.se wrote: ...and I'm sorry about that, the following documents are somewhat contradictory: http://gluster.org/community/documentation/index.php/Simplified_dev_workflow The script will ask you to enter a bugzilla bug id. Every change submitted to GlusterFS needs a bugzilla entry to be accepted. If you do not already have a bug id, file a new bug at Red Hat Bugzilla. If the patch is submitted for review, the rfc.sh script will return the gerrit url for the review request. www.gluster.org/community/documentation/index.php/Development_Work_Flow Prompt for a Bug Id for each commit (if it was not already provded) and include it as a BUG: tag in the commit log. You can just hit enter at this prompt if your submission is purely for review purposes. /Anders -- Anders Blomdell Email: anders.blomd...@control.lth.se Department of Automatic Control Lund University Phone:+46 46 222 4625 P.O. Box 118 Fax: +46 46 138118 SE-221 00 Lund, Sweden ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] glustershd status
This is a small memory system like 1024M and a disk space for the volume is 9gig, i do not think it has anything to do with AFR per se - same bug is also reproducible on the bricks, nfs server too. Also it might be that we aren't able to capture glusterdumps on non Linux platforms properly - one of reasons i used Valgrind output. In Valgrind it indicates about 'lost memory' blocks - You can see the screenshots too which indicate memory ramp ups in seconds with no i/o, in-fact no data on the volume. The work-around i have seen to contain this issue is to disable self-heal-deamon and NFS - after that the memory remains proper. On an interesting observation after running Gluster management daemon in debugging more - i can see that RPCLNT_CONNECT event() is constantly being triggered - which should only occur once?? per process notification? On Thu, Jul 17, 2014 at 3:38 AM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, I don't actively work on AFR, so I might have missed some things. I looked for the following things in the statedump for any memory allocation related oddities. 1) grep pool-misses *dump* This tells us if there were any objects whose allocated mem-pool wasn't sufficient for the load it was working under. I see that the pool-misses were zero, which means we are doing good with the mem-pools we allocated. 2) grep hot-count *dump* This tells us the no. of objects of any kind that is 'active' in the process while the state-dump was taken. This should allow us to see if the numbers we see are explicable. I see the maximum hot-count across statedumps of processes is 50, which isn't alarming or pointing any obvious memory leaks. The above observations indicate that some object that is not mem-pool allocated is being leaked. Hope this helps, Krish - Original Message - Here you go KP - https://bugzilla.redhat.com/show_bug.cgi?id=1120570 On Thu, Jul 17, 2014 at 12:37 AM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, In addition to the valgrind output, statedump output of glustershd process when the leak is observed would be really helpful. thanks, Krish - Original Message - Nope spoke too early, using poll() has no effect on the memory usage on Linux, so actually back to FreeBSD. On Thu, Jul 17, 2014 at 12:07 AM, Harshavardhana har...@harshavardhana.net wrote: KP, I do have a 3.2Gigs worth of valgrind output which indicates this issue, trying to reproduce this on Linux. My hunch says that 'compiling' with --disable-epoll might actually trigger this issue on Linux too. Will update here once i have done that testing. On Wed, Jul 16, 2014 at 11:44 PM, Krishnan Parthasarathi kpart...@redhat.com wrote: Emmanuel, Could you take statedump* of the glustershd process when it has leaked enough memory to be able to observe and share the output? This might give us what kind of objects are we allocating abnormally high. * statedump of a glusterfs process #kill -USR1 pid of process HTH, Krish - Original Message - On Wed, Jul 16, 2014 at 11:32:06PM -0700, Harshavardhana wrote: On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! I observed memory leaks on long terme usage. My favourite test case is building NetBSD on a replicated/distributed volume, and I can see processes growing a lot during the build. I reported it some time ago, and some leaks were plugged, but obviosuly some remain. valgrind was never ported to NetBSD, hence I lack investigative tools, but I bet the leaks exist on FreeBSD and Linux as well. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
[Gluster-devel] release-3.6 branch created
Hi All, A new branch, 'release-3.6', has been branched from this commit in master: commit 950f9d8abe714708ca62b86f304e7417127e1132 Author: Jeff Darcy jda...@redhat.com Date: Tue Jul 8 21:56:04 2014 -0400 dht: fix rename race You can checkout this branch through: $git checkout -b release-3.6 origin/release-3.6 rfc.sh is being updated to send patches to the appropriate branch. The plan is to have all 3.6.x releases happen off this branch. If you need any fix to be part of a 3.4.x release, please send out a backport of the same from master to release-3.4 after it has been accepted in master. More notes on backporting are available at [1]. The branch can be tracked from git.gluster.org and github. forge.gluster.org will be updated to mirror this branch in a bit. Thanks, Vijay [1] http://www.gluster.org/community/documentation/index.php/Backport_Guidelines ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
[Gluster-devel] What's the impact of enabling the profiler?
What impact, if any, does starting profiling (gluster volume profile $vol start) have on performance? ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Change DEFAULT_WORKDIR from hard-coded value
Anyone? On Mon, Jul 14, 2014 at 6:44 PM, Harshavardhana har...@harshavardhana.net wrote: http://review.gluster.org/#/c/8246/ Two important things it achieves - Break-way from '/var/lib/glusterd' hard-coded previously, instead rely on 'configure' value from 'localstatedir' - Provide 's/lib/db' as default working directory for gluster management daemon for BSD and Darwin based installations ${localstatedir}/db was used for FreeBSD (In fact FreeNAS) - where we are planning for a 3.6 release integration as an experimental option perhaps to begin with. Since ${localstatedir}/lib is non-existent on non-linux platforms, it was decided as a natural choice. Future changes in next set would be migrate all the 'tests/*' to handle non '/var/{lib,db}' directories. Need your reviews on the present patchset, please chime in - thank you! -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Updating blog entry list on www.gluster.org?
- Original Message - | From: Eco Willson ewill...@redhat.com | To: Justin Clift jus...@gluster.org | Cc: Gluster Devel gluster-devel@gluster.org, Humble Chirammal hchir...@redhat.com | Sent: Thursday, July 17, 2014 5:32:13 AM | Subject: Re: [Gluster-devel] Updating blog entry list on www.gluster.org? | | This was supposed to be happening automatically, I pinged Garrett LeSage who | implemented the function to look into the issue. | Thanks Eco! Afaict, the same goes for Forge Activity as well. --Humble | | - Original Message - | From: Justin Clift jus...@gluster.org | To: Eco Willson ewill...@redhat.com | Cc: Gluster Devel gluster-devel@gluster.org, Humble Chirammal | hchir...@redhat.com | Sent: Monday, July 14, 2014 6:03:48 AM | Subject: [Gluster-devel] Updating blog entry list on www.gluster.org? | | Hi Eco, | | Humble just published a blog post about the new GlusterFS 3.4.5 | beta2 RPMs: | | http://blog.gluster.org/2014/07/glusterfs-3-4-5beta2-rpms-are-available-now/ | | How do we get the www.gluster.org website updated, so new blog | posts appear automagically? | | Regards and best wishes, | | Justin Clift | | -- | GlusterFS - http://www.gluster.org | | An open source, distributed file system scaling to several | petabytes, and handling thousands of clients. | | My personal twitter: twitter.com/realjustinclift | | ___ | Gluster-devel mailing list | Gluster-devel@gluster.org | http://supercolony.gluster.org/mailman/listinfo/gluster-devel | ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] What's the impact of enabling the profiler?
On 07/18/2014 03:05 AM, Joe Julian wrote: What impact, if any, does starting profiling (gluster volume profile $vol start) have on performance? Joe, According to the code the only extra things it does is calling gettimeofday() call at the beginning and end of the FOP to calculate latency, increment some variables. So I guess not much? Pranith ___ Gluster-users mailing list gluster-us...@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] Gluster Community meeting starting in 15 minutes
On 17/07/2014, at 7:38 PM, Vijay Bellur wrote: snip Had a discussion with Pranith and we felt that 3.5.2 beta is of more importance than 3.6 community test days. Definitely agree personally. People have 3.5.x in production, so to my thinking issues with that should receive priority. 3.6 will arrive whenever its ready. + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] glustershd status
Harsha, I haven't gotten around looking at the valgrind output. I am not sure if I will be able to do it soon since I am travelling next week. Are you seeing an equal no. of disconnect messages in glusterd logs? What is the ip:port you observe in the RPC_CLNT_CONNECT messages? Could you attach the logs to the bug? thanks, Krish - Original Message - This is a small memory system like 1024M and a disk space for the volume is 9gig, i do not think it has anything to do with AFR per se - same bug is also reproducible on the bricks, nfs server too. Also it might be that we aren't able to capture glusterdumps on non Linux platforms properly - one of reasons i used Valgrind output. In Valgrind it indicates about 'lost memory' blocks - You can see the screenshots too which indicate memory ramp ups in seconds with no i/o, in-fact no data on the volume. The work-around i have seen to contain this issue is to disable self-heal-deamon and NFS - after that the memory remains proper. On an interesting observation after running Gluster management daemon in debugging more - i can see that RPCLNT_CONNECT event() is constantly being triggered - which should only occur once?? per process notification? On Thu, Jul 17, 2014 at 3:38 AM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, I don't actively work on AFR, so I might have missed some things. I looked for the following things in the statedump for any memory allocation related oddities. 1) grep pool-misses *dump* This tells us if there were any objects whose allocated mem-pool wasn't sufficient for the load it was working under. I see that the pool-misses were zero, which means we are doing good with the mem-pools we allocated. 2) grep hot-count *dump* This tells us the no. of objects of any kind that is 'active' in the process while the state-dump was taken. This should allow us to see if the numbers we see are explicable. I see the maximum hot-count across statedumps of processes is 50, which isn't alarming or pointing any obvious memory leaks. The above observations indicate that some object that is not mem-pool allocated is being leaked. Hope this helps, Krish - Original Message - Here you go KP - https://bugzilla.redhat.com/show_bug.cgi?id=1120570 On Thu, Jul 17, 2014 at 12:37 AM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, In addition to the valgrind output, statedump output of glustershd process when the leak is observed would be really helpful. thanks, Krish - Original Message - Nope spoke too early, using poll() has no effect on the memory usage on Linux, so actually back to FreeBSD. On Thu, Jul 17, 2014 at 12:07 AM, Harshavardhana har...@harshavardhana.net wrote: KP, I do have a 3.2Gigs worth of valgrind output which indicates this issue, trying to reproduce this on Linux. My hunch says that 'compiling' with --disable-epoll might actually trigger this issue on Linux too. Will update here once i have done that testing. On Wed, Jul 16, 2014 at 11:44 PM, Krishnan Parthasarathi kpart...@redhat.com wrote: Emmanuel, Could you take statedump* of the glustershd process when it has leaked enough memory to be able to observe and share the output? This might give us what kind of objects are we allocating abnormally high. * statedump of a glusterfs process #kill -USR1 pid of process HTH, Krish - Original Message - On Wed, Jul 16, 2014 at 11:32:06PM -0700, Harshavardhana wrote: On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! I observed memory leaks on long terme usage. My favourite test case is building NetBSD on a replicated/distributed volume, and I can see processes growing a lot during the build. I reported it some time ago, and some leaks were plugged, but obviosuly some remain. valgrind was never ported to NetBSD, hence I lack investigative tools, but I bet the leaks exist on FreeBSD and Linux as well. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes -- Religious confuse piety with mere ritual, the virtuous confuse regulation with outcomes --
Re: [Gluster-devel] glustershd status
Sure will do that! - if i get any clues i might send out a patch :-) On Thu, Jul 17, 2014 at 9:05 PM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, I haven't gotten around looking at the valgrind output. I am not sure if I will be able to do it soon since I am travelling next week. Are you seeing an equal no. of disconnect messages in glusterd logs? What is the ip:port you observe in the RPC_CLNT_CONNECT messages? Could you attach the logs to the bug? thanks, Krish - Original Message - This is a small memory system like 1024M and a disk space for the volume is 9gig, i do not think it has anything to do with AFR per se - same bug is also reproducible on the bricks, nfs server too. Also it might be that we aren't able to capture glusterdumps on non Linux platforms properly - one of reasons i used Valgrind output. In Valgrind it indicates about 'lost memory' blocks - You can see the screenshots too which indicate memory ramp ups in seconds with no i/o, in-fact no data on the volume. The work-around i have seen to contain this issue is to disable self-heal-deamon and NFS - after that the memory remains proper. On an interesting observation after running Gluster management daemon in debugging more - i can see that RPCLNT_CONNECT event() is constantly being triggered - which should only occur once?? per process notification? On Thu, Jul 17, 2014 at 3:38 AM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, I don't actively work on AFR, so I might have missed some things. I looked for the following things in the statedump for any memory allocation related oddities. 1) grep pool-misses *dump* This tells us if there were any objects whose allocated mem-pool wasn't sufficient for the load it was working under. I see that the pool-misses were zero, which means we are doing good with the mem-pools we allocated. 2) grep hot-count *dump* This tells us the no. of objects of any kind that is 'active' in the process while the state-dump was taken. This should allow us to see if the numbers we see are explicable. I see the maximum hot-count across statedumps of processes is 50, which isn't alarming or pointing any obvious memory leaks. The above observations indicate that some object that is not mem-pool allocated is being leaked. Hope this helps, Krish - Original Message - Here you go KP - https://bugzilla.redhat.com/show_bug.cgi?id=1120570 On Thu, Jul 17, 2014 at 12:37 AM, Krishnan Parthasarathi kpart...@redhat.com wrote: Harsha, In addition to the valgrind output, statedump output of glustershd process when the leak is observed would be really helpful. thanks, Krish - Original Message - Nope spoke too early, using poll() has no effect on the memory usage on Linux, so actually back to FreeBSD. On Thu, Jul 17, 2014 at 12:07 AM, Harshavardhana har...@harshavardhana.net wrote: KP, I do have a 3.2Gigs worth of valgrind output which indicates this issue, trying to reproduce this on Linux. My hunch says that 'compiling' with --disable-epoll might actually trigger this issue on Linux too. Will update here once i have done that testing. On Wed, Jul 16, 2014 at 11:44 PM, Krishnan Parthasarathi kpart...@redhat.com wrote: Emmanuel, Could you take statedump* of the glustershd process when it has leaked enough memory to be able to observe and share the output? This might give us what kind of objects are we allocating abnormally high. * statedump of a glusterfs process #kill -USR1 pid of process HTH, Krish - Original Message - On Wed, Jul 16, 2014 at 11:32:06PM -0700, Harshavardhana wrote: On a side note while looking into this issue - I uncovered a memory leak too which after successful registration with glusterd, Self-heal daemon and NFS server are killed by FreeBSD memory manager. Have you observed any memory leaks? I have the valgrind output and it clearly indicates of large memory leaks - perhaps it could be just FreeBSD thing! I observed memory leaks on long terme usage. My favourite test case is building NetBSD on a replicated/distributed volume, and I can see processes growing a lot during the build. I reported it some time ago, and some leaks were plugged, but obviosuly some remain. valgrind was never ported to NetBSD, hence I lack investigative tools, but I bet the leaks exist on FreeBSD and Linux as well. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel -- Religious confuse piety with mere ritual, the virtuous confuse
Re: [Gluster-devel] Inspiration for improving our contributor documentation
On 07/17/2014 07:25 PM, Kaushal M wrote: I came across mediawiki's developer documentation and guides when browsing. These docs felt really good to me, and easy to approach. I feel that we should take inspiration from them and start enhancing our docs. (Outright copying with modifications as necessary, could work too. But that just doesn't feel right) Any volunteers? (I'll start as soon as I finish with the developer documentation for data structures for the components I volunteered earlier) ~kaushal [0] - https://www.mediawiki.org/wiki/Developer_hub I love the idea but not sure about the implementation. i.e. considering we already started with .md pages, why not have same kind of pages as .md files in /doc of gluster? We can modify the README in our project so that people can browse all the details in github? Please let me know your thoughts. Pranith [1] - https://www.mediawiki.org/wiki/Category:New_contributors [2] - https://www.mediawiki.org/wiki/Gerrit/Code_review [3] - https://www.mediawiki.org/wiki/Gerrit [4] - https://www.mediawiki.org/wiki/Gerrit/Tutorial [5] - https://www.mediawiki.org/wiki/Gerrit/Getting_started [6] - https://www.mediawiki.org/wiki/Gerrit/Advanced_usage ... and lots more. ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Inspiration for improving our contributor documentation
On Fri, Jul 18, 2014 at 11:11 AM, Pranith Kumar Karampuri pkara...@redhat.com wrote: On 07/17/2014 07:25 PM, Kaushal M wrote: I came across mediawiki's developer documentation and guides when browsing. These docs felt really good to me, and easy to approach. I feel that we should take inspiration from them and start enhancing our docs. (Outright copying with modifications as necessary, could work too. But that just doesn't feel right) Any volunteers? (I'll start as soon as I finish with the developer documentation for data structures for the components I volunteered earlier) ~kaushal [0] - https://www.mediawiki.org/wiki/Developer_hub I love the idea but not sure about the implementation. i.e. considering we already started with .md pages, why not have same kind of pages as .md files in /doc of gluster? We can modify the README in our project so that people can browse all the details in github? Please let me know your thoughts. These kinds of docs need to indexable and searchable by search engines. Only then will they be useful. I don't think markdown files in the source would be good place for these. The other docs related to source/code documentation can be provided in the source as we are attempting to provide now. These need to be directly accessible for devs when developing, so having them in the git repo is good. Pranith [1] - https://www.mediawiki.org/wiki/Category:New_contributors [2] - https://www.mediawiki.org/wiki/Gerrit/Code_review [3] - https://www.mediawiki.org/wiki/Gerrit [4] - https://www.mediawiki.org/wiki/Gerrit/Tutorial [5] - https://www.mediawiki.org/wiki/Gerrit/Getting_started [6] - https://www.mediawiki.org/wiki/Gerrit/Advanced_usage ... and lots more. ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel