[Gluster-devel] [Fwd: bug report glusterfs crash]
-- == Hitesh Chellani Tel: 510-354-6801 Mobile: 510-754-3258 www.zresearch.com Commoditizing Supercomputing and Superstorage! == ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Plans for FreeBSD support?
Avati, The idea is to be disciplined and stick to the tasks on hand, once we've completed *all tasks related to 1.3 stable release,* we will can re-evaluate the Roadmap and re-caliberate resources for FreeBSD. For now please focus on 1.3 stable release. Thanks Hitesh Hitesh wrote: Avati, *I do not want an time spent in development work on FreeBSD !!! *We are only to support anyone trying to port GlusterFS on FreeBSD. Please convey this unequivocally to everyone on the team; also, Vikas has been given this task of supporting FreeBSD port. Hitesh Anand Avati wrote: Attila, Thanks again for your initial work on helping porting glusterfs to FreeBSD. Based on your help we have been able to get glusterfsd to work on FreeBSD and export volumes successfully. We are willing to provide you more active support in getting glusterfs ported. At the time of your previous patch submission there was a transition in codebase from glusterfs--mainline--2.4 to glusterfs--mainline--2.5, and your patches were on 2.4, where related code were being changed in glusterfs--mainline--2.5, hence I couldnt apply those patches immediately. Apologies for the delay. If you can give a shot in getting the client running as well on FreeBSD, we will all appretiate it a lot! thanks again, avati 2007/5/11, Attila Nagy [EMAIL PROTECTED]: Hello, I would like to reply to the following e-mail (I've just subscribed, so this will be a new thread, sorry): http://lists.nongnu.org/archive/html/gluster-devel/2007-05/msg00113.html First of all, if you need a (or more than one) FreeBSD developer machine(s), just tell me, I can give access to them (it's possible that installing FreeBSD on a local machine can be easier). BTW, I've tried to compile it. I have a FreeBSD/amd64 6-STABLE machine and glusterfs-1.3.0-pre3 (tell me if it's not the right version to start with). So far I did the following: - gunzip/untar - -ldl cleaning from configure - ./configure --disable-ibverbs LDFLAGS=-L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include LIBS=-lpthread CC=gcc42 CXX=g++42 (I use gcc42 because I saw on the webpage, that at least gcc 4.1 should be used when developing) - make CFLAGS=-L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include -DFUSE_USE_VERSION=26 CC=gcc42 CXX=g++42 (it's possible that FUSE_USE_VERSION was only needed for 1.2.3, I can't remember, I started with that first, but thought that I should use the newer one) The compilation first fails at: gcc -DPACKAGE_NAME=\glusterfs\ -DPACKAGE_TARNAME=\glusterfs\ -DPACKAGE_VERSION=\1.3.0-pre3\ -DPACKAGE_STRING=\glusterfs 1.3.0-pre3\ -DPACKAGE_BUGREPORT=\[EMAIL PROTECTED] -DPACKAGE=\glusterfs\ -DVERSION=\1.3.0-pre3\ -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=1 -DYYTEXT_POINTER=1 -DHAVE_LIBPTHREAD=1 -DHAVE_LIBDL=1 -DHAVE_LIBFUSE=1 -DHAVE_BACKTRACE=1 -I. -I. -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -DXLATORDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/xlator\ -DSCHEDULERDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/scheduler\ -DTRANSPORTDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/transport\ -DYY_NO_UNPUT -fPIC -Wall -g -shared -nostartfiles -rdynamic -L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include -MT libglusterfs_la-spec.lex.lo -MD -MP -MF .deps/libglusterfs_la-spec.lex.Tpo -c spec.lex.c -fPIC -DPIC -o .libs/libglusterfs_la-spec.lex.o ./spec.l: In function `yylex': ./spec.l:41: error: `yylval' undeclared (first use in this function) ./spec.l:41: error: (Each undeclared identifier is reported only once ./spec.l:41: error: for each function it appears in.) ./spec.l:41: warning: cast from pointer to integer of different size *** Error code 1 Stop in /src/glusterfs-1.3.0-pre3/libglusterfs/src. *** Error code 1 Stop in /src/glusterfs-1.3.0-pre3/libglusterfs. *** Error code 1 I've hacked the following into libglusterfs/src/spec.l (YYSTYPE lines): %{ #include xlator.h #include y.tab.h #include string.h #define YYSTYPE char * extern YYSTYPE yylval; %} Retry. Now it stops at: gcc -DPACKAGE_NAME=\glusterfs\ -DPACKAGE_TARNAME=\glusterfs\ -DPACKAGE_VERSION=\1.3.0-pre3\ -DPACKAGE_STRING=\glusterfs 1.3.0-pre3\ -DPACKAGE_BUGREPORT=\[EMAIL PROTECTED] -DPACKAGE=\glusterfs\ -DVERSION=\1.3.0-pre3\ -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=1 -DYYTEXT_POINTER=1 -DHAVE_LIBPTHREAD=1 -DHAVE_LIBDL=1 -DHAVE_LIBFUSE=1 -DHAVE_BACKTRACE=1 -I. -I. -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -DXLATORDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/xlator\ -DSCHEDULERDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/scheduler\ -DTRANSPORTDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/transport\ -DYY_NO_UNPUT -fPIC -Wall -g -shared -nostartfiles -rdynamic -L/usr/local/lib -L/usr
Re: [Gluster-devel] Plans for FreeBSD support?
Avati, *I do not want an time spent in development work on FreeBSD !!! *We are only to support anyone trying to port GlusterFS on FreeBSD. Please convey this unequivocally to everyone on the team; also, Vikas has been given this task of supporting FreeBSD port. Hitesh Anand Avati wrote: Attila, Thanks again for your initial work on helping porting glusterfs to FreeBSD. Based on your help we have been able to get glusterfsd to work on FreeBSD and export volumes successfully. We are willing to provide you more active support in getting glusterfs ported. At the time of your previous patch submission there was a transition in codebase from glusterfs--mainline--2.4 to glusterfs--mainline--2.5, and your patches were on 2.4, where related code were being changed in glusterfs--mainline--2.5, hence I couldnt apply those patches immediately. Apologies for the delay. If you can give a shot in getting the client running as well on FreeBSD, we will all appretiate it a lot! thanks again, avati 2007/5/11, Attila Nagy [EMAIL PROTECTED]: Hello, I would like to reply to the following e-mail (I've just subscribed, so this will be a new thread, sorry): http://lists.nongnu.org/archive/html/gluster-devel/2007-05/msg00113.html First of all, if you need a (or more than one) FreeBSD developer machine(s), just tell me, I can give access to them (it's possible that installing FreeBSD on a local machine can be easier). BTW, I've tried to compile it. I have a FreeBSD/amd64 6-STABLE machine and glusterfs-1.3.0-pre3 (tell me if it's not the right version to start with). So far I did the following: - gunzip/untar - -ldl cleaning from configure - ./configure --disable-ibverbs LDFLAGS=-L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include LIBS=-lpthread CC=gcc42 CXX=g++42 (I use gcc42 because I saw on the webpage, that at least gcc 4.1 should be used when developing) - make CFLAGS=-L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include -DFUSE_USE_VERSION=26 CC=gcc42 CXX=g++42 (it's possible that FUSE_USE_VERSION was only needed for 1.2.3, I can't remember, I started with that first, but thought that I should use the newer one) The compilation first fails at: gcc -DPACKAGE_NAME=\glusterfs\ -DPACKAGE_TARNAME=\glusterfs\ -DPACKAGE_VERSION=\1.3.0-pre3\ -DPACKAGE_STRING=\glusterfs 1.3.0-pre3\ -DPACKAGE_BUGREPORT=\[EMAIL PROTECTED] -DPACKAGE=\glusterfs\ -DVERSION=\1.3.0-pre3\ -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=1 -DYYTEXT_POINTER=1 -DHAVE_LIBPTHREAD=1 -DHAVE_LIBDL=1 -DHAVE_LIBFUSE=1 -DHAVE_BACKTRACE=1 -I. -I. -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -DXLATORDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/xlator\ -DSCHEDULERDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/scheduler\ -DTRANSPORTDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/transport\ -DYY_NO_UNPUT -fPIC -Wall -g -shared -nostartfiles -rdynamic -L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include -MT libglusterfs_la-spec.lex.lo -MD -MP -MF .deps/libglusterfs_la-spec.lex.Tpo -c spec.lex.c -fPIC -DPIC -o .libs/libglusterfs_la-spec.lex.o ./spec.l: In function `yylex': ./spec.l:41: error: `yylval' undeclared (first use in this function) ./spec.l:41: error: (Each undeclared identifier is reported only once ./spec.l:41: error: for each function it appears in.) ./spec.l:41: warning: cast from pointer to integer of different size *** Error code 1 Stop in /src/glusterfs-1.3.0-pre3/libglusterfs/src. *** Error code 1 Stop in /src/glusterfs-1.3.0-pre3/libglusterfs. *** Error code 1 I've hacked the following into libglusterfs/src/spec.l (YYSTYPE lines): %{ #include xlator.h #include y.tab.h #include string.h #define YYSTYPE char * extern YYSTYPE yylval; %} Retry. Now it stops at: gcc -DPACKAGE_NAME=\glusterfs\ -DPACKAGE_TARNAME=\glusterfs\ -DPACKAGE_VERSION=\1.3.0-pre3\ -DPACKAGE_STRING=\glusterfs 1.3.0-pre3\ -DPACKAGE_BUGREPORT=\[EMAIL PROTECTED] -DPACKAGE=\glusterfs\ -DVERSION=\1.3.0-pre3\ -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=1 -DYYTEXT_POINTER=1 -DHAVE_LIBPTHREAD=1 -DHAVE_LIBDL=1 -DHAVE_LIBFUSE=1 -DHAVE_BACKTRACE=1 -I. -I. -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -DXLATORDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/xlator\ -DSCHEDULERDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/scheduler\ -DTRANSPORTDIR=\/usr/local/lib/glusterfs/1.3.0-pre3/transport\ -DYY_NO_UNPUT -fPIC -Wall -g -shared -nostartfiles -rdynamic -L/usr/local/lib -L/usr/lib -I/usr/include -I/usr/local/include -MT libglusterfs_la-transport.lo -MD -MP -MF .deps/libglusterfs_la-transport.Tpo -c transport.c -fPIC -DPIC -o .libs/libglusterfs_la-transport.o In file included from transport.c:25: logging.h:43: error: syntax error before _gf_log logging.h:46: error: syntax error
Re: [Gluster-devel] Roadmap and support questions.
Hello Francisco and Guten Tag Steffen Wanted to introduce myself and Z Research, Inc. to you since you asked about commercial support; Z Research is the business face of Gluster and I am one of the co-founders and CEO of Z Research, Inc. Z Research was officially formed in June 2005 by AB (Anand Babu) aka rooty who is the CTO and myself with the goal of commoditizing Supercomputing and Superstorage and in the process validating yet another a business model around Free Software, thus evangelizing Free Software and promoting the fact building businesses around Free Software is the way forward. We have our roots in VA Linux Systems and our engineers were the core of the team that deployed the world's second fastest Supercomputer Thunder (Top 500 June 2004); it was a cluster of 4,096 Itanium 2 processors (1,024 x 4-way nodes), 8TB of RAM and 200TB Storage. Essentially, our forte is Clustering and as you can see, we enabled existing File Systems to have clustering ability instead of writing another Clustered File System from scratch. Gluster is a complete clustering stack comprising of components that address building of Clustered File Storage (GlusterFS), HPC Clusters (GlusterHPC), Visualization Cluster (GlusterViz - to be released in the future), Database Clusters (GlusterDB - to be released in the future) and Single System Image (GlusterSSI - to be released in the future). Currently our focus is the storage vertical and we already have customer deployments for GlusterFS - Stanford SEP and a very large Oil Company; GlusterHPC also has been deployed in a number of places. Our customer base also includes the likes of GE and Thermawave. The business model is akin to the Red Hat business model where the customer pays enterprise class support. We also provide full blown professional services around GlusterFS and GlusterHPC right from designing the solution, hardware selection, supervising integration of the hardware (done by our partners Amax Synnex), on-site deployment and customer training. Feel free to write to me if you have any queries or need more information.. Best regards thanks Hitesh == Hitesh Chellani Tel: 510-354-6801 Mobile: 510-754-3258 www.zresearch.com Commoditizing Supercomputing and Superstorage! == Francisco Reyes wrote: Steffen Grunewald writes: Avati asked for suggestions a few days ago (on this list), and there's some traffic on IRC (#gluster) as well. For Avati's request, please consult the mailing list archive (May 1). Will lookup his post. Thanks. Ahum, are you talking about gluster or glusterfs? Two different cups of tea... Glusterfs. The site has a footer that says to write/call to zresearch. For a newbie like me.. that was the first thing I saw. I suppose it's too early to ask for commercial support yet (given that the first code checkin happened less than 9 months ago)... but nobody will stop you to ask for it :-) Can the roadmap be influenced by funding? Is this a mix open source/commercial type of setup or purely open source? ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] error durring write for stripe unify translator
We sleep in shifts ;-) [EMAIL PROTECTED] wrote: That did the trick! It's already working across 6 storage nodes. Thanks for such a quick turn-around, do you ever get to sleep? ;-) Thanks! -Mic Quoting Amar S. Tumballi [EMAIL PROTECTED]: Mic, Thanks for letting us know about the issue. Actually we found that the last moment changes I made in rr scheduler was leading to an infinite loop of glusterfs client :( That error is fixed now. (check the same ftp dir, you have pre2.2 tarball). Also, while in our testing, found that stripe was not complete. Even thats fixed now. You can try the glusterfs with same config file now. Regards, Amar (bulde on #gluster) On Sat, Mar 03, 2007 at 06:51:23PM -0500, [EMAIL PROTECTED] wrote: Before I ask for help let me just say... wow! What an amazing product! This has the potential to shake the SAN market profoundly. I was disappointed in Luster because it made itself sounds like it didn't require a SAN but you folks are strait forward and to the point. Kudos! Now on to the problem: The process for the glusterfs client is spiking to 100% cpu usage and not responding (have to kill -9) whenever I add a stripe or unify translator to the client volume spec. There isn't anything in the client log, but the server logs show: [Mar 03 19:11:01] [ERROR/common-utils.c:52/full_rw()] libglusterfs:full_rw: 0 bytes r/w instead of 113 This only occurs on file writes. I can touch and read files just fine. None of these problems appear when I just mount a remote volume without the translator. I'm using the latest glusterfs-1.3.0-pre2 code on centos with fuse-2.6.3 (had to apply a patch before fuse module loaded) My client volspec is below: volume client0 type protocol/client option transport-type tcp/client option remote-host 192.168.1.201 option remote-port 6996 option remote-subvolume testgl end-volume volume client1 type protocol/client option transport-type tcp/client option remote-host 192.168.1.202 option remote-port 6996 option remote-subvolume testgl end-volume volume stripe type cluster/stripe subvolumes client1 client0 option stripe-size 131072 # 128k end-volume #volume bricks # type cluster/unify # subvolumes client1 client0 # option scheduler rr #end-volume ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel -- == Hitesh Chellani Tel: 510-354-6801 Mobile: 510-754-3285 www.zresearch.com Commoditizing Supercomputing and Superstorage! == ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel
[Gluster-devel] Re: [List-hacking] GlusterFS 1.0-rc2
Hi Guys.Good Job! I shall now spread the work amongst key partners / potential customers that FS is here, have no fear, have a beer!! Gosh, I put W B Yeats to shame! Too late in the night for me, over out! Hitesh == Hitesh Chellani Tel: 510-354-6801 Cell: 510-754-3258 www.zresearch.com Z RESEARCH - Commoditizing Supercomputing! == Amar S. Tumballi wrote: Hi, Gluster Core team is happy to announce the release of GlusterFS v1.0-rc2. The tarball is available at http://gnu.zresearch.com/pub/gluster/glusterfs/glusterfs-1.0.rc2.tar.gz GlusterFS 1.0-rc2 Release Notes: Focus of this release has been mainly on performance (phase-I) and scalability. Most notable improvement is the asynchronous communication framework to allow fine-grained parallelism. This release includes: * 1.x release uses a new asynchronous communication framework to allow fine-grained parallelism. * New 'write-back' translator. * GlusterFS now controls the flow and main loop by getting rid of fuse stack and only use the low-level fuse interface for binding to VFS. This also removes a great number of unwanted getattr lookups, thereby improving speed. * Transport interface is separated out of protocol code and is now loadable module. * Simplified glusterfsd server startup arguments. Few weeks from now you will find 2 more releases which will introduce massive performance boot with the following modules: ib-verbs transport driver (Infiniband low level RDMA), ld-preload GlusterFS client (direct interface, no context switches), scalable locking support, Vector-I/O (zero copying) and extensive POSIX compliance testing. Regards, Gluster Core Team -- http://www.gluster.org (#gluster on irc.gnu.org) ___ list-hacking mailing list [EMAIL PROTECTED] http://zresearch.com/cgi-bin/mailman/listinfo/list-hacking ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Architecture Testing Help
Hi Mehul, I shall have someone in the team respond to you, btw I confused you for your brother who is at Google, we have met at ICC ...if you remember. Take Care Hitesh == Hitesh Chellani Tel: 510-354-6801 Cell: 510-754-3258 www.zresearch.com Z RESEARCH - Commoditizing Supercomputing! == Mehul Shah wrote: Thanks for replying I want to know how Gluster is structured. Yes i want to test that install GlusterHPC how it functions as a cluster. Later on i had one open source application with that i want to test it but before that i want to know that how GLusterHPC works as cluster computing. When the master server is transfering the image extensions to the client node than where does that image get`s store on the client and how client nodes will be functioning in the cluster. If you still want any information regarding my query feel free to ask. Thanking You, Regards Mehul. On 10/7/06, *Hitesh Chellani* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Hi Mehul, I met with you at TiE along with Pankaj Mhatre. Welcome to Gluster, are you still with the Google File System team? BTW, Gluster is a project by Z Research's engineering team. One of the guys from the software team will respond to you but what do you mean by - what exactly is the architecture of Gluster Do you want to know how Gluster is structured or what architectures Glsuter works with? and after i had setup i want to test HPC how will it do the testing. What testing is it that you want to do? Do yo want to test is the hardware you've just installed Gluster on actually functions as a cluster? Do you have any distributed HPC code that you can run on the cluster? Hitesh == Hitesh Chellani Tel: 510-354-6801 Cell: 510-754-3258 www.zresearch.com http://www.zresearch.com Z RESEARCH - Commoditizing Supercomputing! == Mehul Shah wrote: Hi All, I am new to gluster environment. I had setup GlusterHPC on my master node and also image extensions are transfering to client node from master node. Everything went alright with the help of list Thanks for that to Gluster Team. Now Can anybody send me some architecture help, means what exactly is the architecture of Gluster and after i had setup i want to test HPC how will it do the testing. Thanking You, Regards Mehul ___ Gluster-devel mailing list Gluster-devel@nongnu.org mailto:Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] Architecture Testing Help
Hi Mehul, I met with you at TiE along with Pankaj Mhatre. Welcome to Gluster, are you still with the Google File System team? BTW, Gluster is a project by Z Research's engineering team. One of the guys from the software team will respond to you but what do you mean by - what exactly is the architecture of Gluster Do you want to know how Gluster is structured or what architectures Glsuter works with? and after i had setup i want to test HPC how will it do the testing. What testing is it that you want to do? Do yo want to test is the hardware you've just installed Gluster on actually functions as a cluster? Do you have any distributed HPC code that you can run on the cluster? Hitesh == Hitesh Chellani Tel: 510-354-6801 Cell: 510-754-3258 www.zresearch.com Z RESEARCH - Commoditizing Supercomputing! == Mehul Shah wrote: Hi All, I am new to gluster environment. I had setup GlusterHPC on my master node and also image extensions are transfering to client node from master node. Everything went alright with the help of list Thanks for that to Gluster Team. Now Can anybody send me some architecture help, means what exactly is the architecture of Gluster and after i had setup i want to test HPC how will it do the testing. Thanking You, Regards Mehul ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/gluster-devel