Re: [Gluster-users] GlusterFS on a FreeBSD-Server
Stefan Mohn wrote: Hello, I'm sorry to disturb you again. My problem still exists with glusterfs2.0.0. I've tested it with dbench and after one minute glusterfs crashes. I've used the same setting as described below, but now I've used glusterfs2.0.0 release instead. To test the setting i used dbench -D . 5 the last lines of the serverlog were: pending frames: frame : type(1) op(UNLINK) frame : type(1) op(UNLINK) patchset: 7b2e459db65edd302aa12476bc73b3b7a17b1410 signal received: 11 configuration details:db.h 1 dlfcn 1 libpthread 1 spinlock 1 extattr.h 1 st_atimespec.tv_nsec 1 package-string: glusterfs 2.0.0 There has to be a stack-trace following the lines above in the log. Please paste the stack trace also. The full log below doest not have it either. -Shehjar Here are the Server-Logfiles: http://www.physik.tu-berlin.de/~frendor/glusterfs/glusterfsd2.0.0.log (full version) http://www.physik.tu-berlin.de/~frendor/glusterfs/glusterfsd2.0.0_cut.log (short version) Can maybe someone helping me with this issue? Thanks, Stefan Stefan Mohn schrieb: Hello I want to use GlusterFS as a nfs substitute for a FreeBSD Server and Linux clients. The FreeBSD-server should export several directories. One of these directories contains the home-directory for the client-mashines. The problem is that my GlusterFS setup isn't very stable and crashes quite often. It is less stable, when more than one volume is mounted, but even when only one volume is mountet it isn't stable at all. I've tried some different configurations, but now i'm running out if ideas. I hope you can help me. my server: FreeBSD laverne.physik-pool.tu-berlin.de 7.1-RELEASE amd64 I've testet some glusterfs-releases, beginning with Glusterfs 1.4x. I'm now using glusterfs2.0.0rc4 (also testet with rc2, it's still the same) How Glusterfs was build: support packages: bison-2.3_4,1, fusefs-kmod-0.3.9.p1.20080208_4, fusefs-libs-2.7.3 i've used glusterfs-2.0.0rc4: export LDFLAGS=-L/usr/local/lib -liconv export CFLAGS=-O0 -g -DDEBUG -I/usr/local/include/ ./configure make install clean server-vol-file: http://www.physik.tu-berlin.de/~frendor/glusterfs/new_glusterfs_server.vol volume homeklein-dir type storage/posix # POSIX FS translator option directory /scratch/export2/home # Export this directory end-volume volume homeklein type features/locks subvolumes homeklein-dir end-volume volume server type protocol/server option transport-type tcp/server # For TCP/IP transport # option transport-type tcp # For TCP/IP transport subvolumes homeklein # option auth.addr.lscratch.allow 130.149.58.* # Allow access to export volume option auth.addr.homeklein.allow 130.149.58.* # Allow access to export volume # option auth.addr.lcompile.allow 130.149.58.* # Allow access to export volume end-volume client: Linux 2.6.27-11-generic i686 GNU/Linux (it's a Ubuntu 8.10) Glusterfs: packages: libfuse2, fuse-utils it was build as a package with flex, byacc, fakeroot, dh-make from the source glusterfs2.0.0rc4 client-vol-file: http://www.physik.tu-berlin.de/~frendor/glusterfs/new_glusterfs_client_laverne.vol volume homeklein-dir type storage/posix # POSIX FS translator option directory /scratch/export2/home # Export this directory end-volume volume homeklein type features/locks subvolumes homeklein-dir end-volume volume server type protocol/server option transport-type tcp/server # For TCP/IP transport # option transport-type tcp # For TCP/IP transport subvolumes homeklein # option auth.addr.lscratch.allow 130.149.58.* # Allow access to export volume option auth.addr.homeklein.allow 130.149.58.* # Allow access to export volume # option auth.addr.lcompile.allow 130.149.58.* # Allow access to export volume end-volume some logfiles: I've got 2 sets of logfiles at http://www.physik.tu-berlin.de/~frendor/glusterfs The first logfile documents a crash, without any error notification. I've made some readwrite-tests with a client with bonnie++ and read some mails at another client. serverside: http://www.physik.tu-berlin.de/~frendor/glusterfs/glusterfsd.log (73MB) http://www.physik.tu-berlin.de/~frendor/glusterfs/glusterfsd_cut.log (shortend to 1 MB) clientside: http://www.physik.tu-berlin.de/~frendor/glusterfs/glusterfs_client.log the second one was when i shutdown the clients. nothing special was performed. The logfiles give some hints here: from second_glusterfsd_cut.vol 2009-04-23 22:27:08 D [inode.c:336:__inode_retire] homeklein/inode: retiring inode(140733193388032) lru=348/1024 active=56 purge=1 pending frames: frame : type(1) op(LOOKUP) patchset: cb602a1d7d41587c24379cb2636961ab91446f86 + signal received: 11 configuration details:db.h 1 dlfcn 1 libpthread 1 spinlock 1 extattr.h 1 st_atimespec.tv_nsec 1 package-string: glusterfs 2.0.0rc4 server-log: http://www.physik.tu-berlin.de/~frendor/glusterfs/second_glusterfsd_cut.log client-log:
Re: [Gluster-users] About FileSystem
Vahriç Muhtaryan wrote: Hello To All, I’m interesting with glusterfs. Something I do not understand. Docs said that glusterfs do not need to fsck but I’m not sure. if I understood correctly, I’m sharing related servers related directories for example Server 1 : /home/disk_space_1 à which is limited with server disk space where disk_space_1 folder reside and this folder is on node file system which is xfs, ext3, or another Server 2: /home/disk_space_1 à which is limited with server disk space where disk_space_1 folder reside and this folder is on node file system which is xfs, ext3, or another This means always file check will be happen because bit of glusterfs system on Server 1 and , alwyas fsck could be possible ? That is correct but that is a function(..or limitation..) of the local, on-disk file system and not GlusterFS. In fact, if you use the replicate translator you can have multiple copies of your data on different machines so it is possible for you to stop relying on an fsck for data availability. Second I understood that glusterfs on servers file systems and because of this capacity is Server1 + Server 2 + Server n , am I right ? That depends on how you configure it. What you say above, is valid in case you use distribute or stripe translator but is not valid if you use the replicate translator. İs there any way to support iscsi ? Not through GlusterFS. Looks like NFS is supported but NFSv3 and CIFS will be supported next release does it means without samba glusterfs will support CIFS easyly ? Yes. When CIFS support comes in, it will be usable without requiring Samba. Regards Shehjar Thanks Regards Vahric ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] CIFS GlusterFS 2.2 FUSE and WINDOWS
Vahriç Muhtaryan wrote: Hello, Im waiting CIFS support for becase of our environment and have a few questions . · I.s there any deadline for version 2.2 ? 2.1 is roughly due in August so you could estimate 2.2 to be available roughly 3 months after. · I wonder after CIFS support available how we will connect to glusterfs from our windows servers ? FUSE for windows ready ? because I Windows client will be able to connect to GlusterFS through a CIFS front-end. At the server, the CIFS will run inside the GlusterFS server binary so FUSE will not be required. These are, of course, early ideas on how we want to implement it. Things could change when we get down to it. -Shehjar afraid when see this sentence Is it really true that there are no FUSE port for Windows? If anyone knows of one add it to this list, please. I.n sourceforge. Will be any recommendation after CIFS support and version 2.2 released by glusterfs developers ? Regards Vahric ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Glusterfs 2.0 hangs on high load
Ma-ris Ruskulis wrote: Seems now, when autoscaling is off, glusterfs running stable, at least I could not kill it with iozone. Yes, autoscaling is still under consideration as a feature worth having. Please avoid using it for the time being. Thanks Shehjar Maris Ruskulis wrote: Thank You, for reply! As You can see from config, ping-timeout is not set - default is asumed. Now started glusterfs with 8 threads on both server and client (autoscaling switched off). Hardware: *server1:* lspci 00:00.0 Host bridge: Intel Corporation E7505 Memory Controller Hub (rev 03) 00:00.1 Class ff00: Intel Corporation E7505/E7205 Series RAS Controller (rev 03) 00:01.0 PCI bridge: Intel Corporation E7505/E7205 PCI-to-AGP Bridge (rev 03) 00:02.0 PCI bridge: Intel Corporation E7505 Hub Interface B PCI-to-PCI Bridge (rev 03) 00:02.1 Class ff00: Intel Corporation E7505 Hub Interface B PCI-to-PCI Bridge RAS Controller (rev 03) 00:1d.0 USB Controller: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #1 (rev 02) 00:1d.1 USB Controller: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #2 (rev 02) 00:1d.2 USB Controller: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #3 (rev 02) 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev 82) 00:1f.0 ISA bridge: Intel Corporation 82801DB/DBL (ICH4/ICH4-L) LPC Interface Bridge (rev 02) 00:1f.1 IDE interface: Intel Corporation 82801DB (ICH4) IDE Controller (rev 02) 00:1f.3 SMBus: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) SMBus Controller (rev 02) 02:1c.0 PIC: Intel Corporation 82870P2 P64H2 I/OxAPIC (rev 04) 02:1d.0 PCI bridge: Intel Corporation 82870P2 P64H2 Hub PCI Bridge (rev 04) 02:1e.0 PIC: Intel Corporation 82870P2 P64H2 I/OxAPIC (rev 04) 02:1f.0 PCI bridge: Intel Corporation 82870P2 P64H2 Hub PCI Bridge (rev 04) 03:01.0 RAID bus controller: Intel Corporation RAID Controller 04:02.0 Ethernet controller: Intel Corporation 82540EM Gigabit Ethernet Controller (rev 02) 05:02.0 VGA compatible controller: ATI Technologies Inc Rage XL (rev 27) 05:03.0 Ethernet controller: Intel Corporation 82557/8/9/0/1 Ethernet Pro 100 (rev 0d) cat /proc/cpuinfo processor: 0 vendor_id: GenuineIntel cpu family: 15 model: 2 model name: Intel(R) Xeon(TM) CPU 2.40GHz stepping: 5 cpu MHz: 2392.024 cache size: 512 KB physical id: 0 siblings: 2 core id: 0 cpu cores: 1 apicid: 0 initial apicid: 0 fdiv_bug: no hlt_bug: no f00f_bug: no coma_bug: no fpu: yes fpu_exception: yes cpuid level: 2 wp: yes flags: fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe pebs bts cid xtpr bogomips: 4784.04 clflush size: 64 power management: processor: 1 vendor_id: GenuineIntel cpu family: 15 model: 2 model name: Intel(R) Xeon(TM) CPU 2.40GHz stepping: 5 cpu MHz: 2392.024 cache size: 512 KB physical id: 0 siblings: 2 core id: 0 cpu cores: 1 apicid: 1 initial apicid: 1 fdiv_bug: no hlt_bug: no f00f_bug: no coma_bug: no fpu: yes fpu_exception: yes cpuid level: 2 wp: yes flags: fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe pebs bts cid xtpr bogomips: 4784.16 clflush size: 64 power management: *server2:* lspci00:00.0 Host bridge: Intel Corporation E7520 Memory Controller Hub (rev 0c) 00:00.1 Class ff00: Intel Corporation E7525/E7520 Error Reporting Registers (rev 0c) 00:01.0 System peripheral: Intel Corporation E7520 DMA Controller (rev 0c) 00:02.0 PCI bridge: Intel Corporation E7525/E7520/E7320 PCI Express Port A (rev 0c) 00:04.0 PCI bridge: Intel Corporation E7525/E7520 PCI Express Port B (rev 0c) 00:05.0 PCI bridge: Intel Corporation E7520 PCI Express Port B1 (rev 0c) 00:06.0 PCI bridge: Intel Corporation E7520 PCI Express Port C (rev 0c) 00:1d.0 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #1 (rev 02) 00:1d.1 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #2 (rev 02) 00:1d.2 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #3 (rev 02) 00:1d.7 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB2 EHCI Controller (rev 02) 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev c2) 00:1f.0 ISA bridge: Intel Corporation 82801EB/ER (ICH5/ICH5R) LPC Interface Bridge (rev 02) 00:1f.1 IDE interface: Intel Corporation 82801EB/ER (ICH5/ICH5R) IDE Controller (rev 02) 00:1f.3 SMBus: Intel Corporation 82801EB/ER (ICH5/ICH5R) SMBus Controller (rev 02) 01:00.0 PCI bridge: Intel Corporation 6700PXH PCI Express-to-PCI Bridge A (rev 09) 01:00.1 PIC: Intel Corporation 6700/6702PXH I/OxAPIC Interrupt
Re: [Gluster-users] Glusterfs client stops
Paras Fadte wrote: Hi, Does glusetrfs client have any issues with stability ? I ran glusterfs setup for about 4 days and have encountered issue of client suddenly getting stopped causing Transport end-point not connected error message while accessing the mount point. Does this have anything to do with fuse module ? I am using glusterfs2.0.1 version. Awaiting response and thanks in advance Are you using io-threads with auto-scaling turned on? If yes, please add the following option to the io-threads volume section: option max-threads 32 -Shehjar -Paras ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Gluster 2.0.2 locking up issues
Daniel Jordan Bambach wrote: I'm experiencing various locking up issues ranging from Gluster locking up ( 'ls'ing the mount hangs ), to the whole machine locking up under load. My current config is below (two servers, afring) I would love to be able to get to the bottom of this, because it seems very strange that we should see erratic behaviour on such a simple setup. There is approx 12Gb of files, and to stress test (and heal) i run ls -alR on the mount. This will run for a while and eventually lock up Gluster, and occasionally the machine. I have found that in some cases killing Gluster and re-mounting does not solve the problem (in that perhaps both servers have entered a locked state in some way). Im finding it very hard to collect and debug information of any use, as there is no crashlog, no errors in the volume log. Can anyone suggest what I migth be able to do to extract more information as to what is occuring at lock-up time? volume posix type storage/posix option directory /home/export end-volume volume locks type features/locks subvolumes posix end-volume volume brick type performance/io-threads subvolumes locks option autoscaling on option min-threads 8 option max-threads 32 end-volume I see that the max-threads will never exceed 32 which is a reasonable valueand should work fine in most cases but considering some of the other reports we've been getting, could you please try again but without the autoscaling turned on? It is off by default, so you can simply set the number of threads you need by: option thread-count COUNT ...instead of the three option lines above. Thanks Shehjar volume server type protocol/server option transport-type tcp option auth.addr.brick.allow * subvolumes brick end-volume volume latsrv2 type protocol/client option transport-type tcp option remote-host latsrv2 option remote-subvolume brick end-volume volume afr type cluster/replicate subvolumes brick latsrv2 option read-subvolume brick end-volume volume writebehind type performance/write-behind option cache-size 2MB subvolumes afr end-volume volume cache type performance/io-cache option cache-size 32MB option priority *.pyc:4,*.html:3,*.php:2,*:1 option cache-timeout 5 subvolumes writebehind end-volume ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Need a quick answer on Distributed Replicated Storage questions
Liam Slusser wrote: Thanks for the update Anand. Funny you mention unfs3. Just today one of our engineers at work setup unfs3 against our large production gluster cluster and, so far, it has been very good. I just was reading up on your modified booster version and will also give that a try also. I'm looking forward to testing out the modified unfs3 and the native nfs protocol translator! I'd suggest waiting till Jun 23-24 for 2.0.3 to come out. That has all the improvements/fixes to booster and libglusterfsclient, in order to get unfs3 to work properly. -Shehjar Thanks! ls On Thu, Jun 18, 2009 at 1:36 AM, Anand Babu a...@gluster.com mailto:a...@gluster.com wrote: We made good progress with unfs3 integration using booster model. GlusterFS and unfs3 (modified version) will run in single address space using booster library. This feature is scheduled for 2.1. We will try to have a pre-release available soon (in weeks). GlusterFS v2.2 will have a native NFS protocol translator. iSCSI exporting requires mmap support. You can create image files and losetup them as devices. Then it will be possible to export as iSCSI volumes. We just fixed a bug that caused poor mmap write performance. Work is on the way. We will keep you updated. -- Anand Babu Periasamy GPG Key ID: 0x62E15A31 Blog [http://unlocksmith.org] GlusterFS [http://www.gluster.org] GNU/Linux [http://www.gnu.org] Liam Slusser wrote: Jonathan, You can export a Gluster mount via a client with a NFS server however the performance is pretty poor. As far as i know there is no way to export it with iSCSI. Your best option is to use a single/dual Linux/Solaris iscsi server to boot strap all your systems in xenServer and then use Gluster and fuse to mount your /data drive once the system is up and running. liam On Mon, Jun 15, 2009 at 5:15 PM, Jonathan Bayles jbay...@readytechs.com mailto:jbay...@readytechs.com mailto:jbay...@readytechs.com mailto:jbay...@readytechs.com wrote: Hi all, I am attempting to prevent my company from having to buy a SAN to backend our virtualization platform(xenServer). Right now we have a light workload and 4 dell 2950's (6disks, 1 controller each) to leverage against the storage side. I like what I see in regard to the Distributed Replicated Storage where you essentially create a RAID 10 of bricks. This would work very well for me. The question is, how do I serve this storage paradigm to a front end that's expecting an NFS share or an iSCSI target? Does gluster enable me to access the entire cluster from a single IP? Or is it something I could run on a centos cluster (luci and ricci) and use the cluster suite to present the glustered file system in the form of an NFS share? Let me back up and state my needs/assumptions: * A storage cluster with the capacity equal to at least 1 node(assuming all nodes are the same). * I need to be able to lose/take down any one brick in the cluster at any time without a loss of data. * I need more than the throughput of a single server, if not in overall speed, then in width. * I need to be able to add more bricks in and have the expectation of increased storage capacity and throughput. * I need to present the storage as a single entity as an NFS share or a iSCSI target. If there are any existing models out there please point me too them, I don't mind doing the work I just don't want to re-invent the wheel. Thanks in advance for your time and effort, I know what its like to have to answer newbie questions! ___ Gluster-users mailing list Gluster-users@gluster.org mailto:Gluster-users@gluster.org mailto:Gluster-users@gluster.org mailto:Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org mailto:Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] AFR, writebehind, and debug/trace
Barry Jaspan wrote: I just got started with glusterfs. I read the docs over the weekend and today created a simple setup: two servers exporting a brick and one client mounting them with AFR. I am seeing very poor write performance on a dd test, e.g.: time dd if=/dev/zero of=./local-file bs=8192 count=125000 presumably due to a very large number of write operations (because when I increase the blocksize to 64K, the performance increases by 2x). I enabled the writebehind translator but see no improvement. I then enabled a trace translator on both sides of the writebehind and seem to be seeing that write-behind is not batching any of the operations. By batching, if you mean aggregation of smaller requests to be sent as one large request, then, no, write-behind does just that, it writes behind the actual write request. There is no write buffering going on at this point. We've plans of incorporating write buffering in io-cache for the 2.1 release. -Shehjar Server vol file: volume posix type storage/posix option directory /mnt/glusterfsd-export end-volume volume locks type features/locks subvolumes posix end-volume volume brick type performance/io-threads option thread-count 8 subvolumes locks end-volume volume server type protocol/server option transport-type tcp option auth.addr.brick.allow * subvolumes brick end-volume Client vol file: volume remote1 type protocol/client option transport-type tcp option remote-host web-1 option remote-subvolume brick end-volume volume remote2 type protocol/client option transport-type tcp option remote-host web-2 option remote-subvolume brick end-volume volume replicate type cluster/replicate subvolumes remote1 remote2 end-volume volume trace-below type debug/trace subvolumes replicate end-volume volume writebehind type performance/write-behind option cache-size 1MB subvolumes trace-below end-volume volume trace-above type debug/trace subvolumes writebehind end-volume With this configuration, I re-ran by dd test but with only count=100. The log shows: [r...@web-3 glusterfs-mount]# grep trace /var/log/glusterfs/mnt-glusterfs-mount.log | grep above | wc 245 3591 42117 [r...@web-3 glusterfs-mount]# grep trace /var/log/glusterfs/mnt-glusterfs-mount.log | grep below | wc 252 3678 43095 So, there are as many writes to trace-below as trace-above. What am I not understanding? Thanks! Barry ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Is it stable when the process died without vmp umount() ?
Daesung Kim wrote: In my application, many apache processes use apis in libglusterfsclient.so to be a glusterfs client. And vmp is mounted once when the process is initialized. So if I kill the process, vmp list in the library would be cleared. In this case, Can I feel free when the process died without vmp umount? Does it make garbage data which are not destroyed somewhere in the glusterfs servers? Yes. This is not a problem. Once the process using libglusterfsclient dies the connections and related state at the server are also cleaned up. -Shehjar Thanks. ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] HadoopFS-like gluster setup
Peng Zhao wrote: Hi, all, I'm new to gluster, but found it interesting. I want to setup gluster in a way to be similar with HDFS. There is my sample vol-file: volume posix type storage/posix option directory /data1/gluster end-volume volume locks type features/locks subvolumes posix end-volume volume brick type performance/io-threads subvolumes locks end-volume volume server type protocol/server option transport-type tcp option auth.addr.brick.allow * subvolumes brick end-volume volume compute-5-0 type protocol/client option transport-type tcp option remote-host compute-5-0 option remote-subvolume brick end-volume volume compute-5-1 type protocol/client option transport-type tcp option remote-host compute-5-1 option remote-subvolume brick end-volume volume compute-5-2 type protocol/client option transport-type tcp option remote-host compute-5-2 option remote-subvolume brick end-volume volume compute-5-3 type protocol/client option transport-type tcp option remote-host compute-5-3 option remote-subvolume brick end-volume volume compute-5-4 type protocol/client option transport-type tcp option remote-host compute-5-4 option remote-subvolume brick-ns end-volume volume primary type cluster/replicate option local-volume-name primary subvolumes compute-5-0 compute-5-1 end-volume volume secondary type cluster/replicate option local-volume-name secondary subvolumes compute-5-2 compute-5-3 end-volume volume unified type cluster/unify option scheduler rr option local-volume-name unified # do I need this? option namespace compute-5-4 # do I need this? subvolumes primary secondary end-volume volume writebehind type performance/write-behind option cache-size 1MB subvolumes unified end-volume volume cache type performance/io-cache option cache-size 512MB subvolumes writebehind end-volume The glusterd is up running and no error msg in the logs. However, it reports some error when I tried to mount it: [2009-07-01 09:37:36] E [xlator.c:736:xlator_init_rec] xlator: Initialization of volume 'fuse' failed, review your volfile again [2009-07-01 09:37:36] E [glusterfsd.c:498:_xlator_graph_init] glusterfs: initializing translator failed [2009-07-01 09:37:36] E [glusterfsd.c:1191:main] glusterfs: translator initialization failed. exiting I guess it is a very common question. Anyone has any idea? BR, Try generating the log file with log-level set to bug. You can do so by using the -L DEBUG command line parameter. The debug log level will give us a better idea of what exactly is failing. -Shehjar Gnep ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Performance question
Joe Julian wrote: I'm using an unpatched fuse 2.7.4-1 and glusterfs 2.0.2-1 with the following configs and have this result which surprised me: # dd if=/dev/zero of=foo bs=512k count=1024 1024+0 records in 1024+0 records out 536870912 bytes (537 MB) copied, 14.1538 seconds, 37.9 MB/s # dd if=/dev/zero of=foo bs=512k count=1024 1024+0 records in 1024+0 records out 536870912 bytes (537 MB) copied, 24.4553 seconds, 22.0 MB/s Why is it slower if the file exists? Should it be? It is nearly impossible to tell from the difference between these two runs of dd. We can help better if there are more data points for us to look at. Once you have figures averaged from more than a few runs and if you still see a drop, we'd much appreciate if you file a performance bug with your findings at: http://bugs.gluster.com Thanks Shehjar # Servers # volume posix0 type storage/posix option directory /cluster/0 end-volume volume locks0 type features/locks subvolumes posix0 end-volume volume brick0 type performance/io-threads option thread-count 8 subvolumes locks0 end-volume volume posix1 type storage/posix option directory /cluster/1 end-volume volume locks1 type features/locks subvolumes posix1 end-volume volume brick1 type performance/io-threads option thread-count 8 subvolumes locks1 end-volume volume posix2 type storage/posix option directory /cluster/2 end-volume volume locks2 type features/locks subvolumes posix2 end-volume volume brick2 type performance/io-threads option thread-count 8 subvolumes locks2 end-volume volume posix3 type storage/posix option directory /cluster/3 end-volume volume locks3 type features/locks subvolumes posix3 end-volume volume brick3 type performance/io-threads option thread-count 8 subvolumes locks3 end-volume volume server type protocol/server option transport-type tcp subvolumes brick0 brick1 brick2 brick3 option auth.addr.brick0.allow * option auth.addr.brick1.allow * option auth.addr.brick2.allow * option auth.addr.brick3.allow * end-volume Client volume ewcs2_cluster0 type protocol/client option transport-type tcp option remote-host ewcs2.ewcs.com option remote-subvolume brick0 end-volume volume ewcs2_cluster1 type protocol/client option transport-type tcp option remote-host ewcs2.ewcs.com option remote-subvolume brick1 end-volume volume ewcs2_cluster2 type protocol/client option transport-type tcp option remote-host ewcs2.ewcs.com option remote-subvolume brick2 end-volume volume ewcs2_cluster3 type protocol/client option transport-type tcp option remote-host ewcs2.ewcs.com option remote-subvolume brick3 end-volume volume ewcs4_cluster0 type protocol/client option transport-type tcp option remote-host ewcs4.ewcs.com option remote-subvolume brick0 end-volume volume ewcs4_cluster1 type protocol/client option transport-type tcp option remote-host ewcs4.ewcs.com option remote-subvolume brick1 end-volume volume ewcs4_cluster2 type protocol/client option transport-type tcp option remote-host ewcs4.ewcs.com option remote-subvolume brick2 end-volume volume ewcs4_cluster3 type protocol/client option transport-type tcp option remote-host ewcs4.ewcs.com option remote-subvolume brick3 end-volume volume ewcs7_cluster0 type protocol/client option transport-type tcp option remote-host ewcs7.ewcs.com option remote-subvolume brick0 end-volume volume ewcs7_cluster1 type protocol/client option transport-type tcp option remote-host ewcs7.ewcs.com option remote-subvolume brick1 end-volume volume ewcs7_cluster2 type protocol/client option transport-type tcp option remote-host ewcs7.ewcs.com option remote-subvolume brick2 end-volume volume ewcs7_cluster3 type protocol/client option transport-type tcp option remote-host ewcs7.ewcs.com option remote-subvolume brick3 end-volume volume repl1 type cluster/replicate subvolumes ewcs2_cluster0 ewcs4_cluster0 ewcs7_cluster0 end-volume volume repl2 type cluster/replicate subvolumes ewcs2_cluster1 ewcs4_cluster1 ewcs7_cluster1 end-volume volume repl3 type cluster/replicate subvolumes ewcs2_cluster2 ewcs4_cluster2 ewcs7_cluster2 end-volume volume repl4 type cluster/replicate subvolumes ewcs2_cluster3 ewcs4_cluster3 ewcs7_cluster3 end-volume volume distribute type cluster/distribute subvolumes repl1 repl2 repl3 repl4 end-volume volume writebehind type performance/write-behind option aggregate-size 128KB option cache-size 1MB subvolumes distribute end-volume volume ioc type performance/io-cache option cache-size 512MB subvolumes writebehind end-volume ### mount -t glusterfs /etc/glusterfs/glusterfs-client.vol /mnt/gluster ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org
Re: [Gluster-users] Regarding 2.0.4 booster
Sudipto Mukhopadhyay wrote: My mistake I meant booster log is not getting generated. Thanks and regards, Sudipto Hi A few questions. Please see inlined text. -Original Message- From: gluster-users-boun...@gluster.org [mailto:gluster-users-boun...@gluster.org] On Behalf Of Sudipto Mukhopadhyay Sent: Tuesday, July 14, 2009 2:07 PM To: Shehjar Tikoo Cc: av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Hi Shehjar, Thanks for looking into this issue. The file abc.txt is getting created; if you look at the C program; the following line is basically printing out the file handle on stdout: printf (File handle %d\n,fh); But, the booster log is getting generated. Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, July 14, 2009 2:34 AM To: Sudipto Mukhopadhyay Cc: gluster-users@gluster.org; av...@gluster.com Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: Hi, I trying to run some tests w/booster client library and I had to upgrade the glusterfs to version 2.0.4 from 2.0.2 (due to LD_PRELOAD failures of the booster client library, http://git.savannah.gnu.org/cgit/gluster.git/commit/?id=a3ece0caa52ad2ea cf8a8691aaca53295cde972f). Now, w/version 2.0.4 the LD_PRELOAD works OK, but the booster logs are not generating and I am not sure whether booster is working or, not. I have the following booster.conf and client vol specification: $cat /etc/glusterfs/booster.conf /etc/glusterfs/glusterfs-client.vol /mnt/glusterfs glusterfs subvolume=brick,logfile=/tmp/booster.log,loglevel=DEBUG Please confirm whether the above content in booster.conf is on a single line or two separate lines. It should all be on one line for the logfile setting to be associated with /mnt/glusterfs. $cat /etc/glusterfs/glusterfs-client.vol volume brick type protocol/client option transport-type tcp/client # for TCP/IP transport option remote-host 10.16.80.53 # IP address of the server option remote-subvolume afr # name of the remote volume end-volume volume writebehind type performance/write-behind option window-size 1MB subvolumes brick end-volume I have written a simple program to test few system calls and the booster functionality: int main() { int fh; char buffer[65]; fh = open(/mnt/glusterfs/abc.txt,O_WRONLY); if (fh == -1) { fh = creat(/mnt/glusterfs/abc.txt,O_WRONLY); } printf (File handle %d\n,fh); close(fh); return 0; } When I run my program with LD_PRELOAD option, I get the following message: LD_PRELOAD=/usr/local/lib/glusterfs/glusterfs-booster.so /root/fstest/a.out [2009-07-13 16:19:02] E [libglusterfsclient.c:2447:glusterfs_glh_open] libglusterfsclient: path lookup failed for (/abc.txt) Is the above log message the only message that is being output? Were there more lines that resembled this log line that you probably removed for the purpose of this email? Thanks Shehjar File handle 4 And there is no /tmp/booster.log created. The above log is only observed while creating the file (in case abc.txt is not present). Could you please advice on the booster config file and let me know why I am not seeing any booster log? I'll check why the booster log is not getting generated. In the mean time, can you check if the file abc.txt really does not get created on the backend? If open is failing, it looks like the fd 4 is being returned by creat call. -Shehjar Appreciate your help. Thanks and regards, Sudipto ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Regarding 2.0.4 booster
Sudipto Mukhopadhyay wrote: Hi, After modifying the booster.conf, I started getting the logs under /tmp/booster.log but, I see the following error: [2009-07-20 12:24:58] D [booster.c:1025:booster_init] booster: booster is inited [2009-07-20 12:24:58] D [libglusterfsclient.c:1340:libgf_vmp_search_entry] libglusterfsclient: VMP Entry found: /mnt/glusterfs/abc.txt: /mnt/glusterfs/ [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote1: Connected to 10.16.80.53:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote1: Connected to 10.16.80.53:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote2: Connected to 10.16.80.55:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] D [libglusterfsclient-dentry.c:224:__do_path_resolve] libglusterfsclient-dentry: loc-parent(1) already present. sending lookup for 1//abc.txt [2009-07-20 12:24:58] D [libglusterfsclient-dentry.c:245:__do_path_resolve] libglusterfsclient-dentry: resolved path(/abc.txt) till 1(/). sending lookup for remaining path [2009-07-20 12:24:58] D [libglusterfsclient.c:1608:libgf_client_lookup_cbk] libglusterfsclient: 1: (op_num=0) /abc.txt = -1 (No such file or directory) [2009-07-20 12:24:58] E [libglusterfsclient.c:2447:glusterfs_glh_open] libglusterfsclient: path lookup failed for (/abc.txt) Can you please explain the above errors and what is that causing these errors? The other messages are debug and informational messages. If you'd like to reduce the verbosity of the log, please set the loglevel parameter in the booster.conf to ERROR. Regards -Shehjar Thanks and regards, Sudipto -Original Message- From: Sudipto Mukhopadhyay Sent: Monday, July 20, 2009 3:59 PM To: Shehjar Tikoo Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: RE: [Gluster-users] Regarding 2.0.4 booster Hi Shehjar, The contents in booster.conf are in two separate lines. I will try out the test again by merging the contents of booster.conf into one single line after I get my nodes back. Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Thursday, July 16, 2009 10:08 PM To: Sudipto Mukhopadhyay Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: My mistake I meant booster log is not getting generated. Thanks and regards, Sudipto Hi A few questions. Please see inlined text. -Original Message- From: gluster-users-boun...@gluster.org [mailto:gluster-users-boun...@gluster.org] On Behalf Of Sudipto Mukhopadhyay Sent: Tuesday, July 14, 2009 2:07 PM To: Shehjar Tikoo Cc: av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Hi Shehjar, Thanks for looking into this issue. The file abc.txt is getting created; if you look at the C program; the following line is basically printing out the file handle on stdout: printf (File handle %d\n,fh); But, the booster log is getting generated. Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, July 14, 2009 2:34 AM To: Sudipto Mukhopadhyay Cc: gluster-users@gluster.org; av...@gluster.com Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: Hi, I trying to run some tests w/booster client library and I had to upgrade the glusterfs to version 2.0.4 from 2.0.2 (due to LD_PRELOAD failures of the booster client library, http://git.savannah.gnu.org/cgit/gluster.git/commit/?id=a3ece0caa52ad2ea cf8a8691aaca53295cde972f). Now, w/version 2.0.4 the LD_PRELOAD works OK, but the booster logs are not generating and I am not sure whether booster is working or, not. I have the following booster.conf and client vol specification: $cat /etc/glusterfs/booster.conf /etc/glusterfs/glusterfs-client.vol /mnt/glusterfs glusterfs subvolume=brick,logfile=/tmp/booster.log,loglevel=DEBUG Please confirm whether the above content in booster.conf is on a single line or two separate lines. It should all be on one line for the logfile setting to be associated with /mnt/glusterfs. $cat /etc/glusterfs/glusterfs-client.vol volume brick type protocol/client option transport-type tcp/client # for TCP/IP transport option remote-host 10.16.80.53 # IP address of the server option remote-subvolume afr # name of the remote volume end-volume
Re: [Gluster-users] Regarding 2.0.4 booster
Sudipto Mukhopadhyay wrote: Makes sense. Yes I am running the same program. I will be running couple of more tests to verify this. BTW. two more questions on the related topics: 1. How much of performance boost does booster provide? That it provides a performance boost is evident from some of our tests and from reports from users. The exact improvement really depends on various factors such as the volume configuration, the type of translators used, the network setup, etc. 2. does the following http://www.gluster.org/docs/index.php/BoosterConfiguration#Virtual_Mount _Points mean that from the application we always need to use absolute path to the files being operated? That is a good question. I've also considered the need to remove the dependency on absolute paths but the use cases have been limited or none, till now to help me evolve the exact behaviour. Could you please describe what exactly you have in mind? I could take it from there. One approach is to redirect file system operations into GlusterFS when booster sees a string identifier or a token prepended to a path. This token could be specified through the booster.conf file. Since there will be no / prepended to this string, it does not remain an absolute path anymore. Also, since this is a global string identifier for booster, it is very different from a relative path, so can be used from anywhere in the local file system tree, as long as booster knows where to redirect these operations. Thanks Shehjar Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Monday, July 20, 2009 9:46 PM To: Sudipto Mukhopadhyay Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: Hi, After modifying the booster.conf, I started getting the logs under /tmp/booster.log but, I see the following error: [2009-07-20 12:24:58] D [booster.c:1025:booster_init] booster: booster is inited [2009-07-20 12:24:58] D [libglusterfsclient.c:1340:libgf_vmp_search_entry] libglusterfsclient: VMP Entry found: /mnt/glusterfs/abc.txt: /mnt/glusterfs/ [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote1: Connected to 10.16.80.53:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote1: Connected to 10.16.80.53:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote2: Connected to 10.16.80.55:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] D [libglusterfsclient-dentry.c:224:__do_path_resolve] libglusterfsclient-dentry: loc-parent(1) already present. sending lookup for 1//abc.txt [2009-07-20 12:24:58] D [libglusterfsclient-dentry.c:245:__do_path_resolve] libglusterfsclient-dentry: resolved path(/abc.txt) till 1(/). sending lookup for remaining path [2009-07-20 12:24:58] D [libglusterfsclient.c:1608:libgf_client_lookup_cbk] libglusterfsclient: 1: (op_num=0) /abc.txt = -1 (No such file or directory) [2009-07-20 12:24:58] E [libglusterfsclient.c:2447:glusterfs_glh_open] libglusterfsclient: path lookup failed for (/abc.txt) Assuming you're the using the same program as you'd pasted earlier, this last line is an error that says the file was not found in the file system when trying to open it. This is the reason why your program took the file creation branch in the if block. We can be sure about this error if you could confirm, before running your test program, whether the file really did not exist on the backend. Regards -Shehjar Can you please explain the above errors and what is that causing these errors? Thanks and regards, Sudipto -Original Message- From: Sudipto Mukhopadhyay Sent: Monday, July 20, 2009 3:59 PM To: Shehjar Tikoo Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: RE: [Gluster-users] Regarding 2.0.4 booster Hi Shehjar, The contents in booster.conf are in two separate lines. I will try out the test again by merging the contents of booster.conf into one single line after I get my nodes back. Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Thursday, July 16, 2009 10:08 PM To: Sudipto Mukhopadhyay Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: My mistake I meant booster log is not getting generated. Thanks and regards, Sudipto Hi A few questions. Please see inlined text. -Original Message- From: gluster-users-boun...@gluster.org [mailto:gluster-users-boun...@gluster.org] On Behalf Of Sudipto Mukhopadhyay Sent: Tuesday, July 14, 2009 2:07 PM To: Shehjar Tikoo Cc: av...@gluster.com; gluster-users
Re: [Gluster-users] Regarding 2.0.4 booster
Sudipto Mukhopadhyay wrote: This could be tricky as you don't want too lookup too many alternatives!! But, as you are doing LD_PRELOAD, can you not ask the application to specify the paths (I know it's going to be little error prone based on what application supplies) For example: /mnt/glusterfs If the application run dir is /mnt/application then, ../glusterfs is another way to get to the glusterfs. So, for both the paths the booster intercepts the calls... Doing the redirection might be a bit unclean based on the current working directory. Would fit your need if we alias one VMP to another, so that the underlying GlusterFS is accessible through both paths? For eg, a new option in booster.conf that says, vmpalias=/aliased/mount -Shehjar Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Monday, July 20, 2009 11:28 PM To: Sudipto Mukhopadhyay Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: Makes sense. Yes I am running the same program. I will be running couple of more tests to verify this. BTW. two more questions on the related topics: 1. How much of performance boost does booster provide? That it provides a performance boost is evident from some of our tests and from reports from users. The exact improvement really depends on various factors such as the volume configuration, the type of translators used, the network setup, etc. 2. does the following http://www.gluster.org/docs/index.php/BoosterConfiguration#Virtual_Mount _Points mean that from the application we always need to use absolute path to the files being operated? That is a good question. I've also considered the need to remove the dependency on absolute paths but the use cases have been limited or none, till now to help me evolve the exact behaviour. Could you please describe what exactly you have in mind? I could take it from there. One approach is to redirect file system operations into GlusterFS when booster sees a string identifier or a token prepended to a path. This token could be specified through the booster.conf file. Since there will be no / prepended to this string, it does not remain an absolute path anymore. Also, since this is a global string identifier for booster, it is very different from a relative path, so can be used from anywhere in the local file system tree, as long as booster knows where to redirect these operations. Thanks Shehjar Thanks and regards, Sudipto -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Monday, July 20, 2009 9:46 PM To: Sudipto Mukhopadhyay Cc: serv...@gluster.com; av...@gluster.com; gluster-users@gluster.org Subject: Re: [Gluster-users] Regarding 2.0.4 booster Sudipto Mukhopadhyay wrote: Hi, After modifying the booster.conf, I started getting the logs under /tmp/booster.log but, I see the following error: [2009-07-20 12:24:58] D [booster.c:1025:booster_init] booster: booster is inited [2009-07-20 12:24:58] D [libglusterfsclient.c:1340:libgf_vmp_search_entry] libglusterfsclient: VMP Entry found: /mnt/glusterfs/abc.txt: /mnt/glusterfs/ [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote1: Connected to 10.16.80.53:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote1: Connected to 10.16.80.53:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] N [client-protocol.c:5559:client_setvolume_cbk] remote2: Connected to 10.16.80.55:6996, attached to remote volume 'afr'. [2009-07-20 12:24:58] D [libglusterfsclient-dentry.c:224:__do_path_resolve] libglusterfsclient-dentry: loc-parent(1) already present. sending lookup for 1//abc.txt [2009-07-20 12:24:58] D [libglusterfsclient-dentry.c:245:__do_path_resolve] libglusterfsclient-dentry: resolved path(/abc.txt) till 1(/). sending lookup for remaining path [2009-07-20 12:24:58] D [libglusterfsclient.c:1608:libgf_client_lookup_cbk] libglusterfsclient: 1: (op_num=0) /abc.txt = -1 (No such file or directory) [2009-07-20 12:24:58] E [libglusterfsclient.c:2447:glusterfs_glh_open] libglusterfsclient: path lookup failed for (/abc.txt) Assuming you're the using the same program as you'd pasted earlier, this last line is an error that says the file was not found in the file system when trying to open it. This is the reason why your program took the file creation branch in the if block. We can be sure about this error if you could confirm, before running your test program, whether the file really did not exist on the backend. Regards -Shehjar Can you please explain the above errors and what is that causing these errors? Thanks and regards, Sudipto -Original Message- From: Sudipto Mukhopadhyay Sent: Monday, July 20, 2009 3:59 PM To: Shehjar
Re: [Gluster-users] booster unfs with cluster/distribute doesn't work...
Liam Slusser wrote: I've been playing with booster unfs and found that i cannot get it to work with a gluster config that uses cluster/distribute. I am using Gluster 2.0.3... Thanks. I've seen the stale handle errors while using both replicate and distribute. The fixes are in the repo but not part of a release yet. Release 2.0.5 will contain those changes. In the mean time, if you're really interested, you'd check out the repo as: $ git clone git://git.sv.gnu.org/gluster.git ./glusterfs $ cd glusterfs $ git checkout -b release2.0 origin/release-2.0 Also, we've not yet announced it on the list but a customised version of unfs3 is available at: http://ftp.gluster.com/pub/gluster/glusterfs/misc/unfs3/0.5/unfs3-0.9.23booster0.5.tar.gz It has some bug fixes, performance enhancements and work-arounds to improve behaviour with booster. Some documentation is available at: http://www.gluster.org/docs/index.php/Unfs3boosterConfiguration Thanks Shehjar [r...@box01 /]# mount -t nfs store01:/intstore.booster -o wsize=65536,rsize=65536 /mnt/store mount: Stale NFS file handle (just trying it again and sometimes it will mount...) [r...@box01 /]# mount -t nfs store01:/store.booster -o wsize=65536,rsize=65536 /mnt/store [r...@box01 /]# ls /mnt/store data [r...@box01 store]# cd /mnt/store/data -bash: cd: /mnt/store/data/: Stale NFS file handle [r...@box01 /]# cd /mnt/store [r...@box01 store]# cd data -bash: cd: data/: Stale NFS file handle [r...@box01 store]# Sometimes i can get df to show the actual cluster, but most times it gives me nothing. [r...@box01 /]# df -h FilesystemSize Used Avail Use% Mounted on store01:/store.booster 90T 49T 42T 54% /mnt/store [r...@box01 /]# [r...@box01 /]# df -h FilesystemSize Used Avail Use% Mounted on ... store01:/store.booster - - - - /mnt/store However as soon as i remove the cluster/distribute from my gluster client configuration file it works fine. (Missing 2/3 of the files because my gluster cluster has a distribute of 3 volumes per each of the two servers) A strace of unfs during one of the cd commands above outputs: poll([{fd=4, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=21, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=22, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=23, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}], 4, 2000) = 1 ([{fd=22, revents=POLLIN|POLLRDNORM}]) poll([{fd=22, events=POLLIN}], 1, 35000) = 1 ([{fd=22, revents=POLLIN}]) read(22, \200\0\0\230B\307D\234\0\0\0\0\0\0\0\2\0\1\206\243\0\0\0\3\0\0\0\4\0\0\0\1..., 4000) = 156 tgkill(4574, 4576, SIGRT_1) = 0 tgkill(4574, 4575, SIGRT_1) = 0 futex(0x7fff31c7cb20, FUTEX_WAIT_PRIVATE, 1, NULL) = 0 setresgid(-1, 0, -1)= 0 tgkill(4574, 4576, SIGRT_1) = 0 tgkill(4574, 4575, SIGRT_1) = 0 setresuid(-1, 0, -1)= 0 write(22, \200\0\0 B\307D\234\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0F..., 36) = 36 poll([{fd=4, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=21, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=22, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=23, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}], 4, 2000) = 1 ([{fd=22, revents=POLLIN|POLLRDNORM}]) poll([{fd=22, events=POLLIN}], 1, 35000) = 1 ([{fd=22, revents=POLLIN}]) read(22, \200\0\0\230C\307D\234\0\0\0\0\0\0\0\2\0\1\206\243\0\0\0\3\0\0\0\4\0\0\0\1..., 4000) = 156 tgkill(4574, 4576, SIGRT_1) = 0 tgkill(4574, 4575, SIGRT_1) = 0 setresgid(-1, 0, -1)= 0 tgkill(4574, 4576, SIGRT_1) = 0 tgkill(4574, 4575, SIGRT_1) = 0 setresuid(-1, 0, -1)= 0 write(22, \200\0\0 C\307D\234\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0F..., 36) = 36 poll([{fd=4, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=21, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=22, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}, {fd=23, events=POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND}], 4, 2000 unfinished ... With the booster.fstab debug level set a debug, this is all that shows up in the log file: [2009-07-23 02:52:16] D [libglusterfsclient-dentry.c:381:libgf_client_path_lookup] libglusterfsclient: resolved path(/) to 1/1 [2009-07-23 02:52:17] D [libglusterfsclient.c:1340:libgf_vmp_search_entry] libglusterfsclient: VMP Entry found: /store.booster/: /store.booster/ my /etc/booster.conf /home/gluster/apps/glusterfs-2.0.3/etc/glusterfs/liam.conf /store.booster/ glusterfs subvolume=d,logfile=/home/gluster/apps/glusterfs-2.0.3/var/log/glusterfs/d.log,loglevel=DEBUG,attr_timeout=0 my /etc/exports /store.booster myclient(rw,no_root_squash) my client gluster config (liam.conf): volume brick1a type protocol/client option transport-type tcp option remote-host server1 option remote-subvolume brick1a end-volume volume brick1b type protocol/client option transport-type tcp option
Re: [Gluster-users] Gluster 2.0.3 + Apache on CentOS5 performance issue
Somsak Sriprayoonsakul wrote: Ok, I have a chance to run booster over 2.0.4 Have you tried configuring booster with the help doc available at: http://www.gluster.org/docs/index.php/BoosterConfiguration -Shehjar Please find the attach file for my configuration I did configure boost and try simple ls over my Gluster file system. Here's the output of ls -al [r...@compute-0-3 ~]# LD_PRELOAD=/usr/lib64/glusterfs/glusterfs-booster.so ls -l /gluster/www/ ls: /gluster/www/: Invalid argument ls: /gluster/www/members.pantip.com: Invalid argument ls: /gluster/www/cafe.pantip.com: Invalid argument ls: /gluster/www/admin.pantip.com: Invalid argument ls: /gluster/www/www.pantip.com: Invalid argument ls: /gluster/www/passwd3.sql: Invalid argument ls: /gluster/www/passwd.sql: Invalid argument ls: closing directory /gluster/www/: File descriptor in bad state total 129972 drwxr-xr-x 3 root root 8192 May 11 16:13 admin.pantip.com drwxr-xr-x 5 root root 8192 May 18 11:11 cafe.pantip.com drwxr-xr-x 3 root root 8192 May 11 18:48 members.pantip.com -rw-r--r-- 1 root root 66654820 May 18 10:50 passwd3.sql -rw-r--r-- 1 root root 66225769 May 18 10:33 passwd.sql drwxr-xr-x 11 apache apache 8192 May 18 09:47 www.pantip.com [r...@compute-0-3 ~]# [r...@compute-0-3 ~]# LD_PRELOAD=/usr/lib64/glusterfs/glusterfs-booster.so cp /etc/issue /gluster/ [r...@compute-0-3 ~]# LD_PRELOAD=/usr/lib64/glusterfs/glusterfs-booster.so ls -l /gluster/issue ls: /gluster/issue: Invalid argument -rw-r--r-- 1 root root 47 Aug 2 14:57 /gluster/issue [r...@compute-0-3 ~]# LD_PRELOAD=/usr/lib64/glusterfs/glusterfs-booster.so cat /gluster/issue CentOS release 5.3 (Final) Kernel \r on an \m [r...@compute-0-3 ~]# Despite all those errors, output seems to be fine And this is what inside my booster.log [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:55:05] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:56:27] E [libglusterfsclient.c:4194:__glusterfs_stat] libglusterfsclient: path lookup failed for (/hosts) [2009-08-02 14:56:37] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value [2009-08-02 14:57:00] E [libglusterfsclient.c:4194:__glusterfs_stat] libglusterfsclient: path lookup failed for (/issue) [2009-08-02 14:57:07] E [libglusterfsclient.c:2244:glusterfs_getxattr] libglusterfsclient: invalid argument: value Then, I try to LD_PRELOAD apache (prefork). I change the target from /gluster to /usr/home instead (the web application needs it). Then I tried to strace the httpd process and found that httpd crash at the points where httpd tried to read configuration file stored on Gluster volume (bad file descriptor). I will try to move this configuration file some other places and test again. 2009/7/31 Raghavendra G raghaven...@gluster.com Hi, On Thu, Jul 30, 2009 at 11:39 AM, Somsak Sriprayoonsakulsoms...@gmail.com wrote: Thank you very much for you reply At the time we used 2.0.3, and yes we used stock Apache from CentOS. I will try 2.0.4 very soon to see if it's work. For Booster, it seems not working correctly for me. Booster complains a lots of error with plain 'ls' command (but giving the correct output). Also, with Can you mail those errors? booster, Apache process refuse to start. I will try 2.0.4 to see if it improves. If not, I will attach error log next time. logs are very much appreciated. 2009/7/30 Raghavendra G raghaven...@gluster.com Hi Somsak, Sorry for the delayed reply. Below you've mentioned that you've problems with apache and booster. Going forward, Apache over booster will be the preferred approach. Can you tell us what version of glusterfs you are using? And as I can understand you are using apache 2.2, am I correct? regards, - Original Message - From: Liam Slusser lslus...@gmail.com To: Somsak Sriprayoonsakul soms...@gmail.com Cc: gluster-users@gluster.org Sent: Saturday, July 25, 2009 3:46:14 AM GMT +04:00 Abu Dhabi / Muscat Subject: Re: [Gluster-users] Gluster 2.0.3 + Apache on CentOS5 performance issue I haven't tried an apples to apples comparison with Apache+mod_gluster vs Apache+fuse+gluster however i do run both setups. I load tested both setups so to verified it could handle 4x our normal daily
Re: [Gluster-users] Gluster 2.0.3 + Apache on CentOS5 performance issue
Hi I've filed a bug report which you can track at: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=194 Please do add your email to the CC list and also upload the volfiles which you're using with this test. Thanks Shehjar Somsak Sriprayoonsakul wrote: The behavior is still the same ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] double traffic usage since upgrade?
Mark Mielke wrote: Possibly relevant here - At work, we have used a tool which does something similar to booster to accelerate an extremely slow remote file system. It works the same way with LD_PRELOAD, however, it also requires GLIBC to be compiled with --disable-hidden-plt. Reviewing the Internet for similar solutions, will find PlasticFS which also has the same requirement. Recent versions of GLIBC call open() internally without following the regular the regular PLT name resolution model. This increases performance as the PLT indirect lookup model has an expense associated. For example, GLIBC fopen() calls open() directly rather than going through the PLT. So, overriding open() does not intercept calls to fopen()? Is this something the booster developers are aware of? Have they found a way around this, or is it possible that booster is only boosting *some* types of access, and other types of access are still falling through to FUSE? I've asked the developer who wrote out library what he thought of glusterfs/booster not requiring GLIBC with --disable-hidden-plt, and he thinks glusterfs/booster cannot be working (or cannot be intercepting all calls and some calls are leaking through to FUSE). Comments? If some calls were leaking through, this might have the double traffic effect, since FUSE would have its own cache separate from booster? I dont know what a PLT is but I'll attempt to provide some clarity here. It is true, that booster does not support or boost all system calls. We do not require that glibc be built with --disable-hidden-plt for those calls which we do support. For a start, we've aimed at getting apache and unfs3 to work with booster. The functional support for both in booster is complete in 2.0.6 release. For a list of system calls supported by booster, please see: http://www.gluster.org/docs/index.php/BoosterConfiguration There can be applications which need un-boosted syscalls also to be usable over GlusterFS. For such a scenario we have two ways booster can be used. Both approaches are described at the page linked above but in short, you're right in thinking that when the un-supported syscalls are also needed to go over FUSE, we are, as you said, leaking or redirecting calls over the FUSE mount point. That page is a bit long so feel free to ask any questions here. Thanks -Shehjar Cheers, mark On 08/14/2009 01:22 PM, Anand Avati wrote: I've been running 2.0.3 with two backend bricks and a frontend client of mod_gluster/apache 2.2.11+worker for a few weeks now without much issue. Last night i upgraded to 2.0.6 only to find out that mod_gluster has been removed and is recommending to use the booster library - which is fine but i didnt have time to test it last night so i just mounted the whole filesystem with a fuse mount and figured id test the booster config later and then swap. I did try running the 2.0.3 mod_gluster module with the 2.0.6 bricks but apache kept segfaulting (every 10 seconds) and then would spawn another process which would reconnect and keep going. I figured it was dropping a client request every few seconds which is why i went with the fuse mount until i could test the booster library. That would not work, swapping binaries across versions. Well, before with mod_gluster, we would be pushing around 200mbit of web traffic and it would evenly distribute that 200mbit between our two bricks - so server1 would be pushing 100mbit and server2 would be pushing another 100mbit. Basically both inbound from the backend bricks and outbound from apache was basically identical. Except of course if one of the backend glusterd processes died for whatever reason the other remaining brick would take the whole load and its traffic would double as you would expect. Perfect, all was happy. Now using gluster 2.0.6 and fuse both server bricks are pushing the full 200mbit of traffic - so i basically have 400mbit of incoming traffic from the gluster bricks but the same 200mbit of web traffic. I can deal, but i only have a shared gigabit link between my client server and backend bricks and im already eating up basically 50% of that pipe. It is also putting a much larger load on both bricks since i have basically doubled the disk IO time and traffic. Is this a feature? Bug? If I understand correct, 2.0.3 mod_glusterfs = 1x, 2.0.6 fuse = 2x? Can you describe the files being served? (average file size and number of files) Avati ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org
Re: [Gluster-users] booster
Tim Runion - System Administrator wrote: I also wanted to add to this post about booster. I have tried booster with apache 2.2 Which GlusterFS release are you using? Which tool or test against apache results in a core dump? -Shehjar LD_PRELOAD=”/usr/local/lib/glusterfs/glusterfs-booster.so” /usr/local/apache2/bin/httpd Apache does load but I’m getting core dumps. Maybe I don’t have apache compiled correctly with glusterfs? Anyone that is using booster with Apache, please post your configure statement for apache. Thanks, Tim ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS replacement
Stephan von Krawczynski wrote: Hello all, after playing around for some weeks we decided to make some real world tests with glusterfs. Therefore we took a nfs-client and mounted the very same data with glusterfs. The client does some logfile processing every 5 minutes and needs around 3,5 mins runtime in a nfs setup. We found out that it makes no sense to try this setup with gluster replicate as long as we do not have the same performance in a single server setup with glusterfs. So now we have one server mounted (halfway replicate) and would like to tune performance. Does anyone have experience with some simple replacement like that? We had to find out that almost all performance options have exactly zero effect. The only thing that seems to make at least some difference is read-ahead on the server. We end up with around 4,5 - 5,5 minutes runtime of the scripts, which is on the edge as we need something quite below 5 minutes (just like nfs was). Our goal is to maximise performance in this setup and then try a real replication setup with two servers. The load itselfs looks like around 100 scripts starting at one time and processing their data. Any ideas? What nfs server are you using? The in-kernel one? You could try the unfs3booster server, which is the original unfs3 with our modifications for bug fixes and slight performance improvements. It should give better performance in certain cases since it avoids the FUSE bottleneck on the server. For more info, do take a look at this page: http://www.gluster.org/docs/index.php/Unfs3boosterConfiguration When using unfs3booster, please use GlusterFS release 2.0.6 since that has the required changes to make booster work with NFS. -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] crash in __socket_ioq_new
Chetan Ahuja wrote: I posted this on irc (#gluster) a couple of times in the last few days but got no response. Trying my luck here: chetan I'm seeing this crash in 2.0.1 server codebase #4 0x7f6a42848a56 in free () from /lib/libc.so.6 #5 0x7f6a4157ff99 in __socket_ioq_new (this=value optimized out, buf=0x1 Address 0x1 out of ≠, len=value optimized out, vector=0x7f6a3c019db0, count=1006736528, iobref=0x0) at socket.c:313 #6 0x7f6a41581c08 in socket_event_handler (fd=5883, idx=5883, data=0x6, poll_in=-1, poll_out=1116630944, poll_err=0) at socket.c:796 This happened on volume with 4 bricks. Two subvolumes of two replicas each and the final volume distributed over the replicas. There are a large set of clients writing to these volumes. The write patterns are such that all clients are writing into different files. In other words, no two clients are writing simultaneously to the same file at any time. I haven't seen this stack trace on either the mailing list archives or the bugzilla and I don't see any code changes in the relevant code in 2.0.6 either. If the developers would confirm that if it's a known bug and/or has a known workaround... I'd really appreciate that. Hi A bug in this code path was fixed in 2.0.1. For eg, see: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=29 Please confirm if you continue to observe the crash in subsequent releases. Thanks for reporting anyway. -Shehjar Thanks a lot Chetan ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] booster
Stephan von Krawczynski wrote: Hello all, we would like to try a simple booster configuration. Reading the docs we found this: (http://www.gluster.com/community/documentation/index.php/Booster_Configuration_Reference_Page) Still, applications that need to use the old approach, can do so by simply mounting GlusterFS and pointing the application to perform its file system operations over that mount point. In order to utilize the mount-point bypass approach, booster will have to be LD_PRELOAD'ed using the instructions below but without requiring the booster configuration file or setting the environment variables. Does that mean if we provide a classical gluster mount over e.g. /mnt/glusterfs we do not have to specify any GLUSTERFS_BOOSTER_FSTAB in env at all to redirect fs action on /mnt/glusterfs ? Of course we tried but were unsuccessful besides some core dumps (no matter if booster config file provided or not). Yes. That is expected to work. Please try defining the following env variable. It will give us some debugging output to work with. export GLUSTERFS_BOOSTER_LOG=file-pathname If this does not help, lets test again with 2.0.7. The rc1 should be out later today. In the past few days we've fixed a few causes of crash in booster. We only tried to run some bash scripts with preloaded booster... Do you mean the scripts contained commands with LD_PRELOADed booster? Or were you trying to run bash with LD_PRELOADed booster? The second scenario will not work at this point. Thanks -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] libglusterfsclient
David Saez Padros wrote: Hi Is libglusterfsclient a library that oone can use to build applications that read/write to glusterfs file systems directly, bypassing fuse ? if so, were can i find documentation/examples on how to use it ? That is correct. It is like a user-space API. It allows one to build or customize applications to avoid FUSE and kernel FS API. There is no documentation for libglusterfsclient at this point but booster is a user of libglusterfsclient so that is the first place to look into. Here are some more pointers on how to use it. When going through booster, do ignore the booster-specific code related to mapping glusterfs file handles to POSIX integer file descriptors and also the code for reading the FSTAB files. To learn how to instantiate a libglusterfsclient context, see booster/src/booster.c:booster_mount. It shows how to build a glusterfs_init_params_t structure and pass it to glusterfs_mount. Once the context is initialized in libglusterfsclient, you can use the functions with names starting with glusterfs_* or glusterfs_glh_*. The difference between the two types of API is that the first one, i.e. glusterfs_ type does not require the application to pass a handle. A handle here is just an identifier to tell libglusterfsclient which server(s) it should be talking to. The reason this first API does not require a file handle is because libglusterfsclient maintains an internal mapping for you. This mapping is between a path prefix and the corresponding handle. So for eg, if you call glusterfs_mount as: glusterfs_mount (/test/gluster, ipars); where ipars is a glusterfs_init_params_t, any subsequent file system operation through the API that occurs on the /test/gluster path will be mapped to go over the handle stored internally for this path prefix. This of course tells you that libglusterfsclient only supports working with absolute paths. The above approach is preferable when an application needs to use multiple volfiles. In contrast, if you're not interested in the above approach, you use glusterfs_init API to initialize a handle and then use this handle to operate on the files on the servers. In the long run, we'd really prefer applications using booster since that avoids the need to use the custom libglusterfsclient API. Just slip booster under an app and it works. However, the disadvantage is that booster does not at this time support all the system calls an application might need. For a full list, see: http://www.gluster.com/community/documentation/index.php/Booster_Configuration_Reference_Page -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] very low file creation rate with glusterfs -- result updates
Wei Dong wrote: By using booster, I actually avoid being root on the client side. It would be perfect if the servers can also be run by regular users, even if that means that some features have to be deleted. Can someone explain a little bit why the server side must be run by root? There are plenty of reasons why the FUSE approach needs to be run as root but I am sure others more familiar with FUSE can do a far better job of explaining exactly why. With regards to booster, we do not need root since all file system operations are basically being translated by libglusterfsclient into network IO operations so the kernel's file system API is almost completely bypassed. Know that libglusterfsclient is the library used internally by booster. I know that I should not ask for too much when the robustness of the current codebase is the most import issue at the time. I just want to hear a story about that and maybe hack the code myself. Please, dont hesitate to ask any question about Gluster. We'll try to answer as well given the time and other constraints. Thanks -Shehjar - Wei Wei Dong wrote: I think it is fuse that causes the slowness. I ran all experiments with booster enabled and here's the new figure: http://www.cs.princeton.edu/~wdong/gluster/summary-booster.gif . The numbers are MUCH better than NFS in most cases except for the local setting, which is not practically interesting. The interesting thing is that all of a sudden, the deleting rate drop by 4-10 times -- though I don't really care about file deletion. I must say that I'm totally satisfied by the results. - Wei Wei Dong wrote: Hi All, I complained about the low file creation rate with the glusterfs on my cluster weeks ago and Avati suggested I started with a small number of nodes. I finally get sometime to seriously benchmark glusterfs with Bonnie++ today and the results confirms that glusterfs is indeed slow in terms of file creating. My application is to store a large number of ~200KB image files. I use the following bonnie++ command for evaluation (create 10K files of 200KiB each scattered under 100 directories): bonnie++ -d . -s 0 -n 10:20:20:100 Since sequential I/O is not that interesting to me, I only keep the random I/O results. My hardware configuration is 2xquadcore Xeon E5430 2.66GHz, 16GB memory, 4 x Seagate 1500GiB 7200RPM hard drive. The machines are connected with gigabit ethernet. I ran several GlusterFS configurations, each named as N-R-T, where N is the number of replicated volumes aggregated, R is the number of replications and T is number of server side I/O thread. I use one machine to serve one volume so there are NxR servers and one separate client running for each experiment. On the client side, the server volumes are first replicated and then aggregated -- even with 1-1-2 configuration, the single volume is wrapped by a replicate and a distribute translator. To show the overhead of those translators, I also run a simple configuration which is 1-1-2 without the extra replicate distribute translators, and a local configuration which is simple with client server running on the same machine. These configurations are compared to nfs and nfs-local, which is NFS with server and client on the same machine. The GlusterFS volume file templates are attached to the email. The result is at http://www.cs.princeton.edu/~wdong/gluster/summary.gif . The bars/numbers shown are operations/second, so the larger the better. Following are the messages shown by the figure: 1. GlusterFS is doing a exceptionally good job on deleting files, but creates and reads files much slower than both NFS. 2. At least for one node server configuration, network doesn't affects the file creation rate and does affects file read rate. 3. The extra dummy replicate distribute translators lowers file creation rate by almost half. 4. Replication doesn't hurt performance a lot. 5. I'm running only single-threaded benchmark, so it's hard to say about scalability, but adding more servers does helps a little bit even in single-threaded setting. Note that my results are not really that different from http://gluster.com/community/documentation/index.php/GlusterFS_2.0_I/O_Benchmark_Results, where the single node configuration file create rate is about 30/second. I see no reason why GlusterFS has to be that slower than NFS in file creation in single node configuration. I'm wondering if someone here can help me figure out what's wrong in my configuration or what's wrong in the GlusterFS implementation. - Wei Server volume: volume posix type storage/posix option directory /state/partition1/wdong/gluster end-volume volume lock type features/locks subvolumes posix end-volume volume brick type performance/io-threads option thread-count 2 subvolumes lock end-volume volume server type protocol/server option transport-type tcp option
Re: [Gluster-users] problem with booster when multiple volumes are exported per node
Wei Dong wrote: Hi All, I'm experiencing a problem of booster when the server side nodes have more than one volumes exported. The symptom is that when I run ls MOUNT_POINT with booster, I get something like the following: ls: closing directory MOUNT_POINT: File descriptor in bad state. The server configuration file is the following: volume posix0 type storage/posix option directory /state/partition1/gluster end-volume volume lock0 type features/locks subvolumes posix0 end-volume volume brick0 type performance/io-threads option thread-count 2 subvolumes lock0 end-volume volume posix1 type storage/posix option directory /state/partition2/gluster end-volume volume lock1 type features/locks subvolumes posix1 end-volume volume brick1 type performance/io-threads option thread-count 2 subvolumes lock1 end-volume volume server type protocol/server option transport-type tcp option transport.socket.listen-port 7001 option auth.addr.brick0.allow 192.168.99.* option auth.addr.brick1.allow 192.168.99.* subvolumes brick0 brick1 end-volume On the client side, the bricks on the same server are imported separately. The problem only appears when I use booster. Nothing seems to go wrong when I mount GlusterFS. Also everything is find if I only export one brick from each server. There's also no warning or errors in the log file in all cases. Any one has some idea on what's happening? Please post the contents of booster FSTAB file. It'll tell us which subvolume from the client volfile gets used by booster. If the log file is available, do post that also. Thanks -Shehjar - Wei ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] booster
Stephan von Krawczynski wrote: On Mon, 14 Sep 2009 11:40:03 +0530 Shehjar Tikoo shehj...@gluster.com wrote: We only tried to run some bash scripts with preloaded booster... Do you mean the scripts contained commands with LD_PRELOADed booster? Or were you trying to run bash with LD_PRELOADed booster? The second scenario will not work at this point. Thanks -Shehjar Oh, that's bad news. We tried to PRELOAD booster in front of bash (implicit, called the bash-script with LD_PRELOADED). Is this a general problem or a not-yet-implemented feature? A general problem, I'd say. The last time, i.e. when we revamped booster, we tried running with bash but there was some clash with bash internals. We havent done anything special to fix the problem since then because: 1. it requires changes deep inside GlusterFS and; 2. running bash wasnt a very useful scenario when the LD_PRELOAD variable can be added for the bash environment as a whole. For eg. if you just do export LD_PRELOAD=blah on the command line, you can actually have every program started from that shell use booster. -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] libglusterfsclient
David Saez Padros wrote: Hi In the long run, we'd really prefer applications using booster since that avoids the need to use the custom libglusterfsclient API. Just slip booster under an app and it works. However, the disadvantage is that booster does not at this time support all the system calls an application might need. For a full list, see: http://www.gluster.com/community/documentation/index.php/Booster_Configuration_Reference_Page I supoose that simple applications like cp, mv, rm mkdir are fully supported, is that right ? I hope so. ;) I cant comment on such general tools because all our testing focus wrt booster has been on unfs3 and apache, both of which work. Do try these out for us if convenient. We'd really like bug reports about extending booster to support them. Thanks Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Slow unfs3 + booster
Justice London wrote: I am having issues with slow writes to a gluster replicated setup using booster. on the order of about 800KB/sec. When writing a file to a fuse mounted version of the same filesystem I am able to write of course many, many times that speed. Has anyone gotten this to successfully work at this point? If so, with what changes or configs? I have tried both the standard and modified versions of unfs3. Do you have write-behind in the booster volfile? Which release of GlusterFS? With what wsize value are you mounting NFS server at the client? Preferable size is 64KiB but I think Linux client default is 32KiB. Use the -o wsize=65536 with the mount command to increase this value. Please tell me the output of: $ cat /proc/sys/sunrpc/tcp_slot_table_entries The default is 16 on Linux. Try increasing this to say 64 first, and then to 128. Do tell me if this increases the throughput. -Shehjar Justice London E-mail: jlon...@lawinfo.com ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS problem
Hi Due to time constraints, booster has gone untested for the last couple of months. I suggest using unfsd over fuse for the time being. We'll be releasing an alpha of the NFS translator somewhere in March. Let me know if you'd be interested in doing early testing? Thanks -Shehjar Jonas Bulthuis wrote: Hello, I'm using Gluster with cluster/replicate on two servers. On each of these servers I'm exporting the replicated volume through the UNFSv3 booster provided by Gluster. Multiple nfs clients are using these storage servers and most of the time it seems to work fine. However, sometimes the clients give error messages about a 'Stale NFS Handle' when trying to get a directory listing of some directory on the volume (not all directories gave this problem). Yesterday it happened after reinstalling the client machines. All the client machines had the same problem. Rebooting the client machines did not help. Eventually, restarting the UNFSv3 server solved the problem. At least the problem disappeared for now, but, as it happened twice in a short time now, it seems likely that it will occur again. Does anyone have any suggestion on how to permanently solve this problem? This is the nfs booster configuration we're currently using: /etc/glusterfs/cache_acceptation-tcp.vol /nfsexport_acceptation glusterfs subvolume=cache_acceptation,logfile=/usr/local/var/log/glusterfs/booster_acceptation.log,loglevel=DEBUG,attr_timeout=0 Any help will be very much appreciated. Thanks in advance. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] replicating storage for ESX problem
Romans S(c(ugarevs wrote: My configuration: Two node(raid1) cluster based on glusterfs with NFS server. The system is replicating nicely. Changes are synced. The storage is exported over NFS for vmware esx usage. The problem is when I am using vmware ESX to create a VM on the NSF storage, the result is an empty unusable virtual disk. When I export some local ext3 from one of the nodes, correct 8GB VM disk is created. When I reexport the glusterfs mount, I 'm getting unusable VMs(almost zero size). I suppose it has something to do with so called thin provisioning over NFS. Any advice is appreciated! Are you using NFS with the Gluster Storage Platform or is this your customised setup? -Shehjar Regards, Roman. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Using NFS on client side
Aaron Porter wrote: On Tue, Mar 9, 2010 at 2:10 PM, carlopmart carlopm...@gmail.com wrote: Is it possible to export a glusterfs volume mounted on a client via NFS from this client?? You have to use a userspace NFS server (unfs3 seems to work, mostly). Available at: http://ftp.gluster.com/pub/gluster/glusterfs/misc/unfs3/ -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Old story - glusterfs memory usage
Krzysztof Strasburger wrote: On Mon, Mar 29, 2010 at 12:35:32PM +0530, Shehjar Tikoo wrote: Krzysztof Strasburger wrote: On Mon, Mar 29, 2010 at 11:48:44AM +0530, Shehjar Tikoo wrote: This is a known problem. See a previous email on the devel list about it at: http://article.gmane.org/gmane.comp.file-systems.gluster.devel/1469 A bug is filed is at: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=545 For more info on the drop_caches mentioned in that gluster-devel thread, see: http://www.linuxinsight.com/proc_sys_vm_drop_caches.html Do let us know if dropping caches as shown in that thread helps. Dear Shehjar, thank you, but drop_caches really is not the Holy Grail of this problem. It simply does not change anything, to set it to something != 0. Amon Ott figured it out, that forget() is called, but memory is not freed anyway. Of course, just doing drop_caches is not the silver bullet here. In some cases, it needs to be preceded by a sync. What is your experience with that? In any case, if Ott is right, then we might have a memory leak. The best step is to file a bug. Shehjar - I have sent the response to your private mail; sorry for that. Syncing is IMHO irrelevant here, as there are no dirty buffers to be written out to fs. The files are only opened, stated and closed. I filed related bug report in ancient days, when the bug database was hosted on savannah. If I remember correctly, it has been closed in the meantime. Should I file a new report, or find the old one and reopen it (if it is closed)? Krzysztof We'll prefer having the old one re-opened if you can find it, if not, feel free to file a new one. -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] nfs-alpha feedback
Hi Chad I've filed a bug at: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=819 Would you please upload the complete log file from NFS server, GlusterFS server and the volume files for the same to this bug report? Thanks -Shehjar chadr wrote: ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS file locks
There are a couple of things we can do: - Mail us the Glusterfs log files from the NFS server and the glusterfs servers when the lock script fails. Do file a bug if you can. - On the NFS client machine, before you run the mount command, make sure you run the following command. $ rpc.statd - Run the same perl script but this time at the nfs server over the glusterfs mount point, not at the NFS client. If it runs fine, it is probably related to locking over NFS and we'll look at other places to figure it out. -Shehjar Steinmetz, Ian wrote: I'm seeing an issue where I can't lock files on a NFS exported GlusterFS mount. I have two servers connected to each other doing AFR to provide a high available NFS server (mirror the content, one VIP for NFS mounts to clients). Both of the servers have mounted /mnt/glusterfs using GlusterFS with the client pointing to both servers. I then export the filesyste with NFS. I grabbed a quick perl program that tries to lock a file for testing, which fails only on the glusterfs. When I export a normal directory /mnt/test the locking works. Any ideas appreciated. I have a feeling I've implemented the posix/locks option incorrectly. Both servers are running Ubuntu with identical setups, below are relevant configs. r...@gluster01:/mnt/glusterfs# uname -a Linux gluster01 2.6.31-20-generic-pae #58-Ubuntu SMP Fri Mar 12 06:25:51 UTC 2010 i686 GNU/Linux r...@gluster01:/mnt/glusterfs# cat /etc/exports /mnt/glusterfs ip removed for security/25(rw,no_root_squash,no_all_squash,no_subtree_check,sync,insec ure,fsid=10) /mnt/test ip removed for security/25(rw,no_root_squash,no_all_squash,no_subtree_check,sync,insec ure,fsid=11) * I've tried async, rsync, removing all options except FSID. r...@gluster02:/etc/glusterfs# cat glusterfs.vol volume brick1 type protocol/client option transport-type tcp/client option remote-host ip removed for security # IP address of the remote brick option remote-subvolume brick# name of the remote volume end-volume volume brick2 type protocol/client option transport-type tcp/client option remote-host ip removed for security # IP address of the remote brick option remote-subvolume brick# name of the remote volume end-volume volume afr1 type cluster/afr subvolumes brick1 brick2 end-volume volume writebehind type performance/write-behind option window-size 4MB subvolumes afr1 end-volume volume cache type performance/io-cache option cache-size 512MB subvolumes writebehind end-volume volume readahead type performance/read-ahead option page-size 128KB # unit in bytes subvolumes cache end-volume volume iothreads type performance/io-threads option thread-count 4 option cache-size 64MB subvolumes readahead end-volume r...@gluster02:/etc/glusterfs# cat glusterfsd.vol volume posix type storage/posix option directory /data/export end-volume volume locks type features/posix-locks option manditory on # tried with and without this, found in a search of earlier post subvolumes posix end-volume volume brick type performance/io-threads option thread-count 8 subvolumes locks end-volume volume server type protocol/server option transport-type tcp option auth.addr.brick.allow * option auth.addr.brick-ns.allow * option transport.socket.nodelay on option auth.ip.locks.allow * subvolumes brick end-volume * file to test locking... r...@gluster02:/mnt/glusterfs# cat locktest.pl #!/usr/bin/perl use Fcntl qw(:flock); my $lock_file = 'lockfile'; open(LOCKFILE,$lock_file) or die Cannot open $lock_file: $!\n; print Opened file $lock_file\n; flock(LOCKFILE, LOCK_SH) or die Can't get shared lock on $lock_file: $!\n; print Got shared lock on file $lock_file\n; sleep 2; close LOCKFILE; print Closed file $lock_file\n; exit; *Test run from gluster02 using normal NFS mount: r...@gluster02:/# mount ip removed for security:/mnt/test /mnt/test r...@gluster02:/# cd /mnt/test r...@gluster02:/mnt/test# ./locktest.pl Opened file lockfile Got shared lock on file lockfile Closed file lockfile *Test run from gluster02 using gluster exported NFS mount: r...@gluster02:/# mount 74.81.128.17:/mnt/glusterfs /mnt/test r...@gluster02:/# cd /mnt/test r...@gluster02:/mnt/test# ./locktest.pl Opened file lockfile Can't get shared lock on lockfile: No locks available -- Ian Steinmetz Comcast Engineering - Houston 713-375-7866 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS file locks
Steinmetz, Ian wrote: I've turned on debug logging for the server and client of GlusterFS and amended them below. I've replaced the IP addresses with x.x.x and left the last octet for security. It appears I'm able to lock the file when I run the program directly on the gluster mount point, just not when it's mounted via NFS. I checked and rpc.statd is running. One odd thing, when I run the perl locking program directly on the mnt point, it appears to work but spits out the following log message: [2010-05-03 09:16:39] D [read-ahead.c:468:ra_readv] readahead: unexpected offset (4096 != 362) resetting This is nothing to worry about..just debugging output. . . . [glusterfsd server logfile during above testing] r...@gluster02:/var/log/glusterfs# /usr/sbin/glusterfsd -p /var/run/glusterfsd.pid -f /etc/glusterfs/glusterfsd.vol --log-file /var/log/glusterfsd.log --debug -N . . . Given volfile: +--- ---+ 1: volume posix 2: type storage/posix 3: option directory /data/export 4: end-volume 5: 6: volume locks 7: type features/locks 8: option manditory on I'll continue looking into this but in the mean time you could test over NFS again with manditory changed to mandatory. Let me know if that works. Thanks -Shehjar 9: subvolumes posix 10: end-volume ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] DHT translator problem
We need the logs to figure out the exact problem. Please run the glusterfs command while mounting through FUSE with the following command line option: glusterfs -f your-dht-volfile -L TRACE -l /tmp/dhtlog /mnt/gtest Then perform the same operations and email us the dhtlog file, and we'll see whats going on. Thanks -Shehjar Deyan Chepishev wrote: ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] gluster client hang when using iozone
Tomasz Chmielewski wrote: Am 07.06.2010 13:55, Daniel Maher wrote: Any issue what can be wrong here? Neither the client nor the servers produce anything in logs when it happens (I didn't wait for more than 10 minutes though). What distro ? What kernel version ? Hardware specs ? Debian Lenny, 64 bit, 2.6.26 kernel. The specs are more or less high end. I see there is a bug entry describing a similar issue: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=902 but I don't use any NFS/translators here, so not sure if it's the same or not. The bug also points to a different bug in io-cache, which I also use - so I'll try to disable it and see if it changes anything. That comment on io-cache bug is nfs specific and does not come into play when used with FUSE. -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Gluster Volume Crash
In the crash log at http://dpaste.com/213817/, it is clear that the crash happens in the debug/trace translator but in your log file pasted at http://dpaste.com/213489/, I do not see any debug/trace translator configured into the volume file. IOW, the crash log does not correspond to the volume files that you've pasted. What command did you use to start the GlusterFS mount point on the samba server? ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] transport.remote-port is changing on volume restart
Rafael Pappert wrote: Hello List, I'm evaluate gluster platform as a static file backend for a webserver farm. First of all, I have to say thank you to the guys at gluster, you did an awesome job. But there is one really annoying thing, after each restart of a volume in the volume-manager, i have to change the transport.remote-port in the client.vol and remount the volume on all clients. Why is that required? Is there a specific error message that you encounter? If you have the client logs, that will help even more in figuring out the problem. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Glusterfs 3.1 with Ubuntu Lucid 32bit
- Original Message - From: Deadpan110 deadpan...@gmail.com I have posted this to the lists purely to help others - please do not consider any of the following suitable for a production environment and follow these rough instructions at your own risk. 2: glusterfs NFS Obviously make sure you have nfs-common and portmap installed and then mount in the usual way. I found this method had less mem and CPU overheads but locking seemed really bad with some of my services (Dovecot, SVN) and the locks ultimately caused load to spiral out of control. It may have been a misconfiguration on my behalf! Simply using NFS mounting as a read filesystem without the need for locking worked well... but writing large files seemed to lock up the system also (i did not test this with 1024MB of mem and again, it is possibly a configuration on my behalf). If only a single nfs client machine will be running dovecot/svn, you can use the nolock option at mount time to work-around the missing NLM support. -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS export issue with GlusterFS 3.1
Bernard Li wrote: Hi all: I'm trying to setup an NFS export using GlusterFS 3.1. I have setup a replicated volume using the gluster CLI as follows: Volume Name: share Type: Replicate Status: Started Number of Bricks: 5 Transport-type: tcp Bricks: Brick1: gluster01:/export/share Brick2: gluster02:/export/share Brick3: gluster03:/export/share Brick4: gluster04:/export/share Brick5: gluster05:/export/share I then follow the instructions from the following PDF to setup the NFS server: http://download.gluster.com/pub/gluster/glusterfs/qa-releases/nfs-beta/nfs-beta-rc15/GlusterFS_NFS_Beta_RC15_Release_Notes.pdf Regarding this pdf, only the portions which show mount commands and the FAQ section is applicable to 3.1. In 3.1, NFS gets started by default for a volume started with the volume start command. And I got the following errors: [2010-10-26 16:54:16.671148] E [rpc-clnt.c:338:saved_frames_unwind] (--/usr/lib/libgfrpc.so.0(rpc_clnt_notify+0xd4) [0xcacaca] (--/usr/lib/libgfrpc.so.0(rpc_clnt_connection_cleanup+0x65) [0xcaa764] (--/usr/lib/libgfrpc.so.0(saved_frames_destroy+0x22) [0xcaa6f0]))) rpc-clnt: forced unwinding frame type(GF-DUMP) op(DUMP(1)) called at 2010-10-26 16:54:16.670952 [2010-10-26 16:54:16.671177] M [client-handshake.c:849:client_dump_version_cbk] : some error, retry again later [2010-10-26 16:54:16.671198] E [afr-common.c:2598:afr_notify] share: All subvolumes are down. Going offline until atleast one of them comes back up. This is on CentOS 4, 32-bit. GlusterFS servers are not supported on 32 bit systems. -Shehjar BTW, I didn't see any documentation on how to setup the native NFS server in the wiki -- are updated documentations available there? Also, I noticed that /etc/init.d/glusterfsd doesn't exist in the gluster-core RPM package, should it be included? Thanks, Bernard ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Gluster 3.1 and NFS problem
Hi What kind of app/tool was running on the nfs mount point when it hung? Here are some things we can do to debug further. Please restart the NFS server daemon using the TRACE log level, remount NFS and restart the tests. That way I'll be able to zero-in on the operation that results in this bug. To start the nfs server in TRACE log: $ ps ax|grep gluster|grep nfs Kill the process that is listed here. Then use the same command line as in the ps output, except, change the -L and -l arguments so it looks like: glusterfs -f volfile -L TRACE -l /tmp/nfs.log Next: $ dmesg -c /dev/null; to clear dmesg. Remount and run your apps, and watch for not responding message in dmesg. When that happens, please email me /tmp/nfs.log. Thanks -Shehjar M. Vale wrote: HI, I using Gluster 3.1 and after 2 day of working it stops the NFS mount with the following error: [2010-10-27 14:59:54.687519] I [client-handshake.c:699:select_server_supported_programs] storage-client-0: Using Program GlusterFS-3.1.0, Num (1298437), Version (310) [2010-10-27 14:59:54.720898] I [client-handshake.c:535:client_setvolume_cbk] storage-client-0: Connected to 192.168.2.2:24009, attached to remote volume '/mnt'. [2010-10-27 15:00:48.271921] E [rpcsvc.c:1249:nfs_rpcsvc_program_actor] nfsrpc: RPC program not available [2010-10-27 15:06:45.571535] C [inode.c:899:inode_path] storage/inode: possible infinite loop detected, forcing break. name=(.) [2010-10-27 15:06:45.571818] C [inode.c:899:inode_path] storage/inode: possible infinite loop detected, forcing break. name=((null)) [2010-10-27 15:06:45.572129] C [inode.c:899:inode_path] storage/inode: possible infinite loop detected, forcing break. name=((null)) If a mount the gluster NFS on another machine everything works ok, but if I try to cpoy large data to the NFS mount the mount disapears, and the client says: nfs: server IP not responding, still trying What is this error : torage/inode: possible infinite loop detected, forcing break. name=(.) ? Regards Mauro V. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS export issue with GlusterFS 3.1
Bernard Li wrote: Hi Shehjar: The hosts with exports in the replication pool are already running NFS servers so I need to setup GlusterFS native NFS server on another server. I am using the following /etc/glusterfs/glusterfsd.vol: Thats not recommended but I can see why this is needed. The simplest way to run the nfs server for the two replicas is to simply copy over the nfs volume file from the current nfs server. It will work right away. The volume file below will not. Performance will also drop because now both your replicas are another network hop away. I guess the ideal situation would be to allow gnfs to run even when there is already a server running. Its on the ToDo list. -Shehjar volume gluster01 type protocol/client option transport-type tcp option remote-host gluster01 option remote-port 6996 option remote-subvolume brick end-volume volume gluster02 type protocol/client option transport-type tcp option remote-host gluster02 option remote-port 6996 option remote-subvolume brick end-volume volume repl type cluster/replicate subvolumes gluster01 gluster02 end-volume volume writebehind type performance/write-behind option cache-size 4MB subvolumes repl end-volume volume dshare type performance/io-cache option cache-size 1GB subvolumes writebehind end-volume volume nfs-server type nfs/server subvolumes dshare end-volume and I start glusterfsd Does this look about right? Is the remote-port correct? Thanks, Bernard On Tue, Oct 26, 2010 at 10:14 PM, Shehjar Tikoo shehj...@gluster.com wrote: Bernard Li wrote: On Tue, Oct 26, 2010 at 9:15 PM, Shehjar Tikoo shehj...@gluster.com wrote: Regarding this pdf, only the portions which show mount commands and the FAQ section is applicable to 3.1. In 3.1, NFS gets started by default for a volume started with the volume start command. So basically you're saying if I have a 2 server replicated volume on gluster01 and gluster02, and I did volume start on gluster01, I should be able to mount the volume via NFS from either of the servers? Yes. You can test that both servers have exported the same volume by: $ showmount -e gluster01 $ showmount -e gluster02 Is exporting via glusterfsd still supported? Yes. -Shehjar Cheers, Bernard ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] cannot nfs mount glusterFS
Rick King wrote: Craig, Can you let us know the bug number? I ran into this as well. Heres the one specific to the nfs problem above. http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1639 ~~Rick - Original Message - From: Craig Carl cr...@gluster.com To: Matt Hodson ma...@geospiza.com Cc: gluster-users@gluster.org Sent: Wednesday, November 3, 2010 12:41:28 PM Subject: Re: [Gluster-users] cannot nfs mount glusterFS There is a bug filed, Gluster should throw a warning when you start the volume. Please keep us updated as you test, let me know f you have any other questions. Thanks, Craig -- Craig Carl Gluster, Inc. Cell - (408) 829-9953 (California, USA) Gtalk - craig.c...@gmail.com From: Matt Hodson ma...@geospiza.com To: Vikas Gorur vi...@gluster.com Cc: gluster-users@gluster.org Sent: Wednesday, November 3, 2010 11:34:29 AM Subject: Re: [Gluster-users] cannot nfs mount glusterFS HA! that was it. dolt! thank you. i was going crazy looking at other stuff. -matt --- Matt Hodson Scientific Customer Support, Geospiza (206) 633-4403, Ext. 111 http://www.geospiza.com On Nov 3, 2010, at 11:26 AM, Vikas Gorur wrote: On Nov 3, 2010, at 11:18 AM, Matt Hodson wrote: I just installed distributed gluster FS on 2 CentOS 5 boxes. install and configuration seemed to go fine. gluterd is running. firewalls/iptables are off. however for the life of me i cannot nfs mount the main gluster server from either a OSX or a CentOS 5 box. I use NFS often and have a fair amount of experience with it so i've reviewed most of the common pitfalls. here's the command that fails from centos: $ sudo mount -v -t nfs 172.16.1.76:/gs-test /mnt/gluster/ mount: trying 172.16.1.76 prog 13 vers 3 prot tcp port 2049 mount: trying 172.16.1.76 prog 15 vers 3 prot udp port 909 mount: 172.16.1.76:/gs-test failed, reason given by server: No such file or directory and the same one from OSX 10.5 sudo mount -v -t nfs 172.16.1.76:/gs-test /gluster/ mount_nfs: can't access /gs-test: No such file or directory what's weird is that i can mount actual dirs on the gluster server, just not the gluster VOLNMAE. in other words, this command works fine because it's mounting an actual dir. $sudo mount -v -t nfs 172.16.1.76:/ /mnt/gluster/ You have the kernel NFS service running. That is why you can mount regular directories on the gluster server. When you try to mount Gluster the kernel NFS server is actually looking for a directory called /gs-test, which ofcourse does not exist. You need to stop the kernel NFS service and stop and start the gluster volume. -- Vikas Gorur Engineer - Gluster, Inc. -- ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users DISCLAIMER: This e-mail and any files transmitted with it ('Message') is intended only for the use of the recepient (s) named and may contain confidential information. Opinions, conclusion and other information in this message that do not relate to the official business of King7. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Cannot mount NFS
Please try some of the steps mentioned at: http://www.gluster.org/faq/index.php?sid=679lang=enaction=showcat=5 Thanks Horacio Sanson wrote: I have a two server replicated volume that I can mount without problems using the native client but I cannot make this work via NFS no matter what. I read the mailing list and the FAQ, applied all fixes but still no luck. System: Two Gluster nodes with vanilla Ubuntu 10.10 LTS 64bit. One client with vanilla Ubuntu 10.10 Desktop 32bit One client with vanilla Ubuntu 10.10 LTS 64bit Gluster: Installed using glusterfs_3.1.0-1_amd64.deb Command used to create volume: sudo gluster volume create www replica 2 transport tcp \ 192.168.4.90:/opt/www 192.168.4.91:/opt/www Below I present all my attempts to get NFS mounted on the two clients I have so everything below multiply by 2: 1. NFS Mount attempt 1: After creating the volume $ sudo mount -v -t nfs 192.168.4.91:/www /mnt mount.nfs: timeout set for Thu Nov 4 10:37:17 2010 mount.nfs: trying text-based options 'vers=4,addr=192.168.4.91,clientaddr=192.168.4.111' mount.nfs: mount(2): Connection refused 2. NFS Mount attempt 2: Seems UDP is not supported so I added the tcp option: $ sudo mount -v -o mountproto=tcp -t nfs 192.168.4.91:/www /mnt mount.nfs: timeout set for Thu Nov 4 10:38:36 2010 mount.nfs: trying text-based options 'mountproto=tcp,addr=192.168.4.91' mount.nfs: prog 13, trying vers=3, prot=6 mount.nfs: portmap query retrying: RPC: Unable to receive mount.nfs: prog 13, trying vers=3, prot=17 mount.nfs: portmap query failed: RPC: Unable to receive - Connection refused mount.nfs: trying text-based options 'mountproto=tcp,addr=192.168.4.91' mount.nfs: prog 13, trying vers=3, prot=6 mount.nfs: portmap query retrying: RPC: Unable to receive 3. NFS Mount attempt 3: Google tells me I need to start portmap - on both Gluster servers and the clients I installed portmap $ sudo aptitude install portmap # This should be a dependency of Gluster deb $ sudo service portmap start - on the client: $ sudo mount -v -o mountproto=tcp -t nfs 192.168.4.91:/www /mnt mount.nfs: timeout set for Thu Nov 4 10:42:07 2010 mount.nfs: trying text-based options 'mountproto=tcp,addr=192.168.4.91' mount.nfs: prog 13, trying vers=3, prot=6 mount.nfs: portmap query retrying: RPC: Program not registered mount.nfs: prog 13, trying vers=3, prot=17 mount.nfs: portmap query failed: RPC: Program not registered mount.nfs: prog 13, trying vers=2, prot=6 mount.nfs: portmap query retrying: RPC: Program not registered mount.nfs: prog 13, trying vers=2, prot=17 mount.nfs: portmap query failed: RPC: Program not registered mount.nfs: prog 13, trying vers=2, prot=6 mount.nfs: portmap query retrying: RPC: Program not registered mount.nfs: prog 13, trying vers=2, prot=17 mount.nfs: portmap query failed: RPC: Program not registered mount.nfs: requested NFS version or transport protocol is not supported 4. Me throwing keyboard throught the fourth floor window. 5. NFS Mount attempt 4: Retry 3 but without tcp option, maybe it is not needed with portmap started: $ sudo mount -v -t nfs 192.168.4.91:/www /mnt mount.nfs: timeout set for Thu Nov 4 10:45:37 2010 mount.nfs: trying text-based options 'vers=4,addr=192.168.4.91,clientaddr=192.168.4.111' mount.nfs: mount(2): Connection refused mount.nfs: trying text-based options 'vers=4,addr=192.168.4.91,clientaddr=192.168.4.111' mount.nfs: mount(2): Connection refused mount.nfs: trying text-based options 'vers=4,addr=192.168.4.91,clientaddr=192.168.4.111' mount.nfs: mount(2): Connection refused 5. Re-read FAQ, there is something about DNS being use for authentication. Sounds related to the Connection refused error I am getting: skil...@store90:~$ sudo gluster set www rpc-auth.addr.namelookup off unrecognized word: set (position 0) skil...@store90:~$ sudo gluster set www rpc-auth.addr.namelookup off unrecognized word: set (position 0) skil...@store90:~$ sudo gluster set www rpc-auth.addr.namelookup off unrecognized word: set (position 0) skil...@store90:~$ sudo gluster set www namelookup off unrecognized word: set (position 0) skil...@store90:~$ sudo gluster set www add.namelookup off unrecognized word: set (position 0) skil...@store90:~$ sudo gluster set www rpc-auth.namelookup off unrecognized word: set (position 0) skil...@store90:~$ sudo gluster set www rpc-auth.addr.namelookup off unrecognized word: set (position 0) 6. Set up hosts file so each Gluster node and client can resolve their hostnames locally. $ NFS Mount attempt 5,6,7,8: Try all mount options from 1-5 above, including 4 several times but with the hosts file correctly set up. 7. Bang head against wall get huge cup of coffee. 8. Re-read mailing list: It seems that gluster NFS may conflict with the kernel NFS service: $ sudo aptitude search nfs-kernel-server p nfs-kernel-server it is not installed so this could not be the problem.
Re: [Gluster-users] Gluster crash
Please file a bug. It'd help to have the steps to reproduce and if it is easily reproduced, the client log in TRACE log level. Thanks. Samuel Hassine wrote: Hi all, Our service using GlusterFS is in production since one week and we are managing a huge trafic. The last night, one of the Gluster client (on a physical node with a lot of virtual engines) crashed. Can you give me more information about the log of the crash? Here is the log: pending frames: frame : type(1) op(READ) frame : type(1) op(READ) frame : type(1) op(READ) frame : type(1) op(CREATE) frame : type(1) op(CREATE) patchset: v3.0.6 signal received: 6 time of crash: 2010-11-06 05:38:11 configuration details: argp 1 backtrace 1 dlfcn 1 fdatasync 1 libpthread 1 llistxattr 1 setfsid 1 spinlock 1 epoll.h 1 xattr.h 1 st_atim.tv_nsec 1 package-string: glusterfs 3.0.6 /lib/libc.so.6[0x7f7644e76f60] /lib/libc.so.6(gsignal+0x35)[0x7f7644e76ed5] /lib/libc.so.6(abort+0x183)[0x7f7644e783f3] /lib/libc.so.6(__assert_fail+0xe9)[0x7f7644e6fdc9] /lib/libpthread.so.0(pthread_mutex_lock+0x686)[0x7f76451a0b16] /lib/glusterfs/3.0.6/xlator/performance/io-cache.so(ioc_create_cbk +0x87)[0x7f7643dcd3f7] /lib/glusterfs/3.0.6/xlator/performance/read-ahead.so(ra_create_cbk +0x1a2)[0x7f7643fd9322] /lib/glusterfs/3.0.6/xlator/cluster/replicate.so(afr_create_unwind +0x126)[0x7f76441f1866] /lib/glusterfs/3.0.6/xlator/cluster/replicate.so(afr_create_wind_cbk +0x10f)[0x7f76441f25ef] /lib/glusterfs/3.0.6/xlator/protocol/client.so(client_create_cbk +0x5aa)[0x7f764443a00a] /lib/glusterfs/3.0.6/xlator/protocol/client.so(protocol_client_pollin +0xca)[0x7f76444284ba] /lib/glusterfs/3.0.6/xlator/protocol/client.so(notify +0xe0)[0x7f7644437d70] /lib/libglusterfs.so.0(xlator_notify+0x43)[0x7f76455cd483] /lib/glusterfs/3.0.6/transport/socket.so(socket_event_handler +0xe0)[0x7f76433819e0] /lib/libglusterfs.so.0[0x7f76455e7e0f] /sbin/glusterfs(main+0x82c)[0x40446c] /lib/libc.so.6(__libc_start_main+0xe6)[0x7f7644e631a6] /sbin/glusterfs[0x402a29] I just want to know why Gluster crashed. Regards. Sam ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] question on NFS mounting
Joe Landman wrote: On 11/07/2010 02:00 AM, Bernard Li wrote: I'm not sure about distribute, but with replicate, each brick should be able to act as the NFS server. What does `showmount -e` say for each brick? And what error message did you get when you tried to mount it? With any kind of volume config, NFS starts up by default on all bricks. You'll have ensure that no other nfs servers are running on the bricks when Gluster volumes are started. Actually, showmount didn't work. We get permission denied. Even after playing with the auth.allowed flag. Please paste the output of rpcinfo -p brick-addr. It'll help point out whats going on. Thanks -Shehjar Cheers, Bernard ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS crashes under load
Thanks. I'll be looking into it. I've filed a bug at: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2061 You may add yourself to the CC list for notifications. It seems the crash is easily reproduced on your setup. Can you please post the log from Gluster NFS process in TRACE log level to the bug? Dan Bretherton wrote: I upgraded to GlusterFS 3.1 a couple of weeks ago and overall I am very impressed; I think it is a big step forward. Unfortunately there is one feature that is causing me a big problem - the NFS process crashes every few hours when under load. I have pasted the relevant error messages from nfs.log at the end of this message. The rest of the log file is swamped with these messages incidentally. [2010-11-06 23:07:04.977055] E [rpcsvc.c:1249:nfs_rpcsvc_program_actor] nfsrpc: RPC program not available There are no apparent problems while these errors are being produced so this issue probably isn't relevant to the crashes. Correct. That error is misleading and will be removed in 3.1.1 Thanks -Shehjar To give an indication of what I mean by under load, we have a small HPC cluster that is used for running ocean models. A typical model run involves 20 processors, all needing to read simultaneously from the same input data files at regular intervals during the run. There are roughly 20 files, each ~1GB in size. At the same time this is going on several people, typically, are processing output from previous runs from this and other (much bigger) clusters, chugging through hundreds of GB and tens of thousands of files every few hours. I don't think the Gluster-NFS crashes are purely load dependant because they seem to occur at different load levels, which is what leads me to suspect something subtle related to the cluster's 20-processor model runs. I would prefer to use the GlusterFS client on the cluster's compute nodes, but unfortunately the pre-FUSE Linux kernel has been customised in a way that has thwarted all my attempts to build a FUSE module that the kernel will accept (see http://gluster.org/pipermail/gluster-users/2010-April/004538.html) The servers that are exporting NFS are all running CentOS 5.5 with GlusterFS installed from RPMs, and the GlusterFS volumes are distributed (not repicated). Two of the servers with GlusterFS bricks are actually running SuSE Enterprise 10; I don't know if this is relevant. I used previous GlusterFS versions with SLES10 without any problems, but as RPMs are not provided for SuSE I presume it is not an officially supported distro. For that reason I am only using the CentOS machines as NFS servers for the GlusterFS volumes. I would be very grateful for any suggested solutions or workarounds that might help to prevent these NFS crashes. -Dan. nfs.log extract -- [2010-11-06 23:07:10.380744] E [fd.c:506:fd_unref_unbind] (--/usr/lib64/glusterfs/3.1.0/xlator/debug/io-stats.so(io_stats_fstat_cbk+0x8e) [0x2b30813e] (--/usr/lib64/glusterfs/3.1.0/xlator/nfs/server.so(nfs_fop_fstat_cbk+0x41) [0x2b9a6da1] (--/usr/lib64/glusterfs/3.1.0/xlator/nfs/server.so(nfs3svc_readdir_fstat_cbk+0x22d) [0x2b9b0bdd]))) : Assertion failed: fd-refcount pending frames: patchset: v3.1.0 signal received: 11 time of crash: 2010-11-06 23:07:10 configuration details: argp 1 backtrace 1 dlfcn 1 fdatasync 1 libpthread 1 llistxattr 1 setfsid 1 spinlock 1 epoll.h 1 xattr.h 1 st_atim.tv_nsec 1 package-string: glusterfs 3.1.0 /lib64/libc.so.6[0x35746302d0] /lib64/libpthread.so.0(pthread_spin_lock+0x2)[0x357520b722] /usr/lib64/libglusterfs.so.0(fd_unref_unbind+0x3d)[0x38f223511d] /usr/lib64/glusterfs/3.1.0/xlator/nfs/server.so(nfs3svc_readdir_fstat_cbk+0x22d)[0x2b9b0bdd] /usr/lib64/glusterfs/3.1.0/xlator/nfs/server.so(nfs_fop_fstat_cbk+0x41)[0x2b9a6da1] /usr/lib64/glusterfs/3.1.0/xlator/debug/io-stats.so(io_stats_fstat_cbk+0x8e)[0x2b30813e] /usr/lib64/libglusterfs.so.0(default_fstat_cbk+0x79)[0x38fa69] /usr/lib64/glusterfs/3.1.0/xlator/performance/read-ahead.so(ra_attr_cbk+0x79)[0x2aeec459] /usr/lib64/glusterfs/3.1.0/xlator/performance/write-behind.so(wb_fstat_cbk+0x9f)[0x2ace402f] /usr/lib64/glusterfs/3.1.0/xlator/cluster/distribute.so(dht_attr_cbk+0xf4)[0x2b521d24] /usr/lib64/glusterfs/3.1.0/xlator/protocol/client.so(client3_1_fstat_cbk+0x287)[0x2aacd2b7] /usr/lib64/libgfrpc.so.0(rpc_clnt_handle_reply+0xa2)[0x38f1a0f2e2] /usr/lib64/libgfrpc.so.0(rpc_clnt_notify+0x8d)[0x38f1a0f4dd] /usr/lib64/libgfrpc.so.0(rpc_transport_notify+0x2c)[0x38f1a0a77c] /usr/lib64/glusterfs/3.1.0/rpc-transport/socket.so(socket_event_poll_in+0x3f)[0x2aaac3eb435f] /usr/lib64/glusterfs/3.1.0/rpc-transport/socket.so(socket_event_handler+0x168)[0x2aaac3eb44e8] /usr/lib64/libglusterfs.so.0[0x38f2236ee7] /usr/sbin/glusterfs(main+0x37d)[0x4046ad] /lib64/libc.so.6(__libc_start_main+0xf4)[0x357461d994] /usr/sbin/glusterfs[0x402dc9] - ___ Gluster-users mailing list
Re: [Gluster-users] question on NFS mounting
Mike Hanby wrote: Each Gluster server that is going to also act as an NFS server (or Samba) has to mount the volume using the Gluster client: There is no mounting taking place to export GlusterFS through NFS. That NFS server is a GlusterFS client, is correct but it is also a translator. That means any translation from NFS ops to GlusterFS ops happens internally in the Gluster NFS server process, without having to go through a mount point. -Shehjar For example, on my two GlusterFS servers I have the following: LABEL=brick01-01/export/nas-01/brick01 ext4defaults,_netdev 0 2 LABEL=brick01-02/export/nas-01/brick02 ext4defaults,_netdev 0 2 ...bricks 3-8 LABEL=brick01-09/export/nas-01/brick09 ext4defaults,_netdev 0 2 localhost:/dev-storage /developglusterfs defaults,_netdev 0 0 LABEL=brick02-01/export/nas-02/brick01 ext4defaults,_netdev 0 2 LABEL=brick02-02/export/nas-02/brick02 ext4defaults,_netdev 0 2 ...bricks 3-8 LABEL=brick02-09/export/nas-02/brick09 ext4defaults,_netdev 0 2 localhost:/dev-storage /developglusterfs defaults,_netdev 0 0 And then /usr/sbin/showmount -e nas-01 and /usr/sbin/showmount -e nas-02 both show the /dev-storage * export. Also verify your firewalls have the appropriate ports open on all servers. For example on CentOS 5 for 2 servers with 18 bricks and 1 volume: # GlusterFS -A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 111 -j ACCEPT -A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 111 -j ACCEPT # GlsuterFS 24007, 24008 plus 1 port per brick across all volumes, 18 bricks in this case -A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 24007:24026 -j ACCEPT # Gluster port 38465 and NFS for 2 Gluster servers, 1 port per server -A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 38465:38468 -j ACCEPT # End GlusterFS Mike -Original Message- From: gluster-users-boun...@gluster.org [mailto:gluster-users-boun...@gluster.org] On Behalf Of Bernard Li Sent: Sunday, November 07, 2010 1:10 AM To: land...@scalableinformatics.com Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] question on NFS mounting Hi Joe: On Sun, Nov 7, 2010 at 12:03 AM, Joe Landman land...@scalableinformatics.com wrote: Actually, showmount didn't work. We get permission denied. Even after playing with the auth.allowed flag. That's an indication that the gNFS server is not running. I would recommend you review the FAQ and some of the recent posts on the list, as there have been a couple threads discussing numerous NFS-related issues and their solutions. I've collected them for you here for your convenience: http://www.gluster.org/faq/index.php?sid=679lang=enaction=showcat=5 http://gluster.org/pipermail/gluster-users/2010-November/005692.html Cheers, Bernard ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] question on NFS mounting
Hi Joe, Was your permission denied problem solved? How? I'd like to add an entry into the NFS FAQ page about this. Thanks Joe Landman wrote: Hi Folks We have a 3.1 cluster set up, and NFS mounting is operational. We are trying to get our heads around the mounting of this cluster. What we found works (for a 6 brick distributed cluster) is using the same server:/export in all the mounts. My questions are 1) can we use any of the bricks for server? We tried using another brick in the volume, but it doesn't seem to work. 2) if we need to use a single server for the mount, is this a performance bottleneck? That is, does all the traffic have to traverse one brick? 3) in previous releases (3.0.x and before) we could mount the file system on each server. We are doing this now with the nfs mount, but my concern from point 1 still stands. We are working on understanding the performance issues with this. Hopefully with some benchmarks soon. Regards, Joe ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] ACL with GlusterFS 3.1?
Hi ACLs are not supported as yet. Thanks Mike Hanby wrote: Howdy, Are access control lists (ACL, i.e. setfacl / getfacl) supported in GlusterFS 3.1? If yes, beyond mounting the bricks with defaults,acl what do I need to do to enable ACL for both NFS and native Gluster clients? Google isn't returning anything useful on this topic. Thanks, Mike = Mike Hanby mha...@uab.edu UAB School of Engineering Information Systems Specialist II IT HPCS / Research Computing ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS Mounted GlusterFS, secondary groups not working
Hi, It might be related to a bug filed at: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2045 If you please update it there or file a new one, I'll take a look. Thanks. - Original Message - From: Mike Hanby mha...@uab.edu To: gluster-users@gluster.org Sent: Friday, November 12, 2010 12:00:23 AM Subject: [Gluster-users] NFS Mounted GlusterFS, secondary groups not working Howdy, I have a GlusterFS 3.1 volume being mounted on a client using NFS. From the client I created a directory under the mount point and set the permissions to root:groupa 750 My user account is a member of groupa on the client, yet I am unable to list the contents of the directory: $ ls -l /gfs/dir1 ls: /gfs/dir1/: Permission denied $ ls -ld /gfs/dir1 rwxr-x--- 9 root groupa 73728 Nov 9 09:44 /gfs/dir1/ $ groups myuser groupa I am able to list the directory as the user root. If I change the group ownership to my primary group, myuser, then I can successfully list the contents of the directory. $ sudo chgrp myuser /gfs/dir1 $ ls -ld /gfs/dir1 rwxr-x--- 9 root myuser 73728 Nov 9 09:44 /gfs/dir1/ $ ls -l /gfs/dir1 drwxr-xr-x 5 root root 73728 Mar 26 2010 testdir1 drwxr-x--- 4 root root 73728 Apr 8 2010 testdir2 drwxr-x--- 2 root root 73728 Aug 4 21:23 testdir3 The volume is being exported using the builtin GlusterFS NFS server. The servers and client are all CentOS 5.5 x86_64 boxes. Thanks for any suggestions, Mike = Mike Hanby mha...@uab.edu UAB School of Engineering Information Systems Specialist II IT HPCS / Research Computing ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Exposing parts of a volume to specific clients?
Mike Hanby wrote: Howdy, We have 18TB at our disposal to share via GlusterFS 3.1. My initial thought was to create a single volume, comprised of 18 x 1TB bricks. The volume will be used for user storage as well as storage for applications. Is there any way to create different exports for the various clients via NFS and Gluster client? For example, the server used by our user base should be able to mount the users directory on the gfs volume (nas-01:/gfs/users), where as the yum repository server should only be able to mount nas-01:/gfs/repo Essentially, I only want to expose what's necessary for each client machine to perform its roll. If this isn't possible directly within GlusterFS, I'm open to suggestions to rethink my strategy. Not possible yet. Its on the ToDo list. Thanks Thanks, Mike ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Exposing parts of a volume to specific clients?
Mike Hanby wrote: So, to accomplish this right now I would need to disable the builtin NFS server, enable the OS NFS server, mount the Gluster file system and export as normal via /etc/exports? Just brainstorming here, You can still export directories individually but the support is minimal as far as restricting the exports to different clients. More info at: http://www.gluster.org/faq/index.php?sid=940lang=enaction=artikelcat=5id=55artlang=en -Shehjar Mike -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Thursday, November 11, 2010 10:47 PM To: Mike Hanby Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] Exposing parts of a volume to specific clients? Mike Hanby wrote: Howdy, We have 18TB at our disposal to share via GlusterFS 3.1. My initial thought was to create a single volume, comprised of 18 x 1TB bricks. The volume will be used for user storage as well as storage for applications. Is there any way to create different exports for the various clients via NFS and Gluster client? For example, the server used by our user base should be able to mount the users directory on the gfs volume (nas-01:/gfs/users), where as the yum repository server should only be able to mount nas-01:/gfs/repo Essentially, I only want to expose what's necessary for each client machine to perform its roll. If this isn't possible directly within GlusterFS, I'm open to suggestions to rethink my strategy. Not possible yet. Its on the ToDo list. Thanks Thanks, Mike ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] XenServer and Glusterfs 3.1
Vikas Gorur vi...@gluster.com wrote: On Dec 8, 2010, at 2:23 PM, Michael Patterson wrote: Hi Shehjar, I'm currently running gluster 3.1.1 on Centos 5.5 x64 and I am unable to mount NFS subdirectories from xenserver. Was support for this feature added in gluster 3.1.1? I've also tried using various options defined here, with no luck: http://gluster.org/faq/index.php?sid=2118lang=enaction=artikelcat=5id=55artlang=en Like Stefano, I'm using the 'option nfs.port 2049*'*' in my /etc/glusterd/nfs/nfs-server.vol configuration. Xenserver can create the SR, but it cannot mount the subdirectory (see below). - # glusterd -V glusterfs 3.1.1 built on Nov 29 2010 10:07:44 Mounting SR subdirectory from xenserver fails: --- # mount 10.0.0.1:/vol0/2e8a14e9-83d6-6e68-7c1a-464e6691988b /mnt mount: 10.0.0.1:/vol0/2e8a14e9-83d6-6e68-7c1a-464e6691988b failed, reason given by server: No such file or directory --- Are you sure you don't have the Linux kernel NFS server running on 10.0.0.1? Make sure you disable it before starting the Gluster volume. You can verify that the Gluster NFS server is running by running this on 10.0.0.1: # showmount -e localhost The output should show: /vol0/2e8a14e9-83d6-6e68-7c1a-464e6691988b * -- Vikas Gorur Engineer - Gluster, Inc. -- Yes. Support for xen server subdirectory mounting is not present yet. It is still a work in progress. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Read-only volumes - timeframe for features/filter?
Are any apps on the mount point erroring out with: Invalid argument or Stale NFS file handle? Burnash, James wrote: Hello. Has anyone seen error messages like this in /var/log/glusterfs/nfs.log: tail /var/log/glusterfs/nfs.log [2011-01-10 14:22:55.859066] I [afr-self-heal-common.c:1526:afr_self_heal_completion_cbk] pfs-ro1-replicate-3: background meta-data data self-heal completed on / [2011-01-10 14:22:55.859084] I [afr-self-heal-common.c:1526:afr_self_heal_completion_cbk] pfs-ro1-replicate-5: background meta-data data self-heal completed on / [2011-01-10 14:22:57.786088] E [nfs3.c:4346:nfs3_fsstat] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:22:59.355112] E [nfs3.c:4346:nfs3_fsstat] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:22:59.415732] E [nfs3.c:675:nfs3_getattr] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:22:59.455029] E [nfs3.c:675:nfs3_getattr] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:23:01.800751] E [nfs3.c:675:nfs3_getattr] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:23:02.127233] E [nfs3.c:675:nfs3_getattr] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:23:07.834044] E [nfs3.c:4346:nfs3_fsstat] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:23:09.478852] E [nfs3.c:4346:nfs3_fsstat] nfs-nfsv3: Failed to map FH to vol [2011-01-10 14:40:18.558072] E [afr-self-heal-metadata.c:524:afr_sh_metadata_fix] pfs-ro1-replicate-5: Unable to self-heal permissions/ownership of '/' (possible split-brain). Please fix the file on all backend volumes Mount is done with this command: mount -v -t nfs -o soft,rsize=16384,wsize=16384 jc1lpfsnfsro:/pfs-ro1 /pfs1 Command line being executed is: rsync -av --progress /pfs1/online_archive/2010 . This CentOS 5.5 x86_64, GlusterFS 3.1.1. Currently configured: gluster volume info Volume Name: pfs-ro1 Type: Distributed-Replicate Status: Started Number of Bricks: 10 x 2 = 20 Transport-type: tcp Bricks: Brick1: jc1letgfs17-pfs1:/export/read-only/g01 Brick2: jc1letgfs18-pfs1:/export/read-only/g01 Brick3: jc1letgfs17-pfs1:/export/read-only/g02 Brick4: jc1letgfs18-pfs1:/export/read-only/g02 Brick5: jc1letgfs17-pfs1:/export/read-only/g03 Brick6: jc1letgfs18-pfs1:/export/read-only/g03 Brick7: jc1letgfs17-pfs1:/export/read-only/g04 Brick8: jc1letgfs18-pfs1:/export/read-only/g04 Brick9: jc1letgfs17-pfs1:/export/read-only/g05 Brick10: jc1letgfs18-pfs1:/export/read-only/g05 Brick11: jc1letgfs17-pfs1:/export/read-only/g06 Brick12: jc1letgfs18-pfs1:/export/read-only/g06 Brick13: jc1letgfs17-pfs1:/export/read-only/g07 Brick14: jc1letgfs18-pfs1:/export/read-only/g07 Brick15: jc1letgfs17-pfs1:/export/read-only/g08 Brick16: jc1letgfs18-pfs1:/export/read-only/g08 Brick17: jc1letgfs17-pfs1:/export/read-only/g09 Brick18: jc1letgfs18-pfs1:/export/read-only/g09 Brick19: jc1letgfs17-pfs1:/export/read-only/g10 Brick20: jc1letgfs18-pfs1:/export/read-only/g10 Thanks. James Burnash Unix Engineering. DISCLAIMER: This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this in error, please immediately notify me and permanently delete the original and any copy of any e-mail and any printout thereof. E-mail transmission cannot be guaranteed to be secure or error-free. The sender therefore does not accept liability for any errors or omissions in the contents of this message which arise as a result of e-mail transmission. NOTICE REGARDING PRIVACY AND CONFIDENTIALITY Knight Capital Group may, at its discretion, monitor and review the content of all e-mail communications. http://www.knight.com ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Write performance on XenServer guest
Shehjar Tikoo wrote: What gluster config are you using? My bad.. I didnt read the mail completely. First thing you should try is run a streaming IO write perf test using dd or iozone. Lets see how that performs over the replicated config. Thanks Stefano Baronio wrote: Hello, I have recently set up gluster-3.1.2 as NFS Virtual Disk Storage for XenServer. I have ran a windows vm on it and tested the disk performance: Read: 100MB/s Write: 10MB/s While with standard NFS, on the same servers, we can achieve: Read: 115MB/s Write: 100MB/s We have two servers with local scsi disks with gluster configured in replicate mode. Volume Options: performance.cache-size 1GB write-behind-window-size 512MB performance.stat-prefech 1 Is there any specific option for this configuration? And, is it normal such a poor write performance? Thank you Stefano Baronio ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Fwd: files not syncing up with glusterfs 3.1.2
Hi Paul, Locking is part of the core GlusterFS protocol but the NFS server module does not have NLM support yet(NLM is the locking protocol associated with NFSv3). On linux, the workaround is generally to mount with the -o nolock option although I dont see why excluding this option results in stale file handle and other errors. let me go through the complete thread, I'll reply elsewhere. Today, if locking among multiple client machines is a must-have, you'll have to use FUSE. paul simpson wrote: so, while your all about - my big question is can/does gluster (with NFS/fuse client) properly lock files? ie, a simple test is to checkout a svn tree to a gluster, modify, checkin, list, alter, revert. everytime i do this with 3.1.2 i get input/output errors from my client machine within a few minutes. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Fwd: files not syncing up with glusterfs 3.1.2
paul simpson wrote: hello all, i have been testing gluster as a central file server for a small animation studio/post production company. my initial experiments were using the fuse glusterfs protocol - but that ran extremely slowly for home dirs and general file sharing. we have since switched to using NFS over glusterfs. NFS has certainly seemed more responsive re. stat and dir traversal. however, i'm now being plagued with three different types of errors: 1/ Stale NFS file handle 2/ input/output errors 3/ and a new one: $ l -l /n/auto/gv1/production/conan/hda/published/OLD/ ls: cannot access /n/auto/gv1/production/conan/hda/published/OLD/shot: Remote I/O error total 0 d? ? ? ? ?? shot ...so it's a bit all over the place. i've tried rebooting both servers and clients. these issues are very erratic - they come and go. some information on my setup: glusterfs 3.1.2 g1:~ # gluster volume info Volume Name: glustervol1 Type: Distributed-Replicate Status: Started Number of Bricks: 4 x 2 = 8 Transport-type: tcp Bricks: Brick1: g1:/mnt/glus1 Brick2: g2:/mnt/glus1 Brick3: g3:/mnt/glus1 Brick4: g4:/mnt/glus1 Brick5: g1:/mnt/glus2 Brick6: g2:/mnt/glus2 Brick7: g3:/mnt/glus2 Brick8: g4:/mnt/glus2 Options Reconfigured: performance.write-behind-window-size: 1mb performance.cache-size: 1gb performance.stat-prefetch: 1 network.ping-timeout: 20 diagnostics.latency-measurement: off diagnostics.dump-fd-stats: on that is 4 servers - serving ~30 clients - 95% linux, 5% mac. all NFS. Mac OS as a nfs client remains untested against Gluster NFS. Do you see these errors on Mac or Linux clients? other points: - i'm automounting using NFS via autofs (with ldap). ie: gus:/glustervol1 on /n/auto/gv1 type nfs (rw,vers=3,rsize=32768,wsize=32768,intr,sloppy,addr=10.0.0.13) gus is pointing to rr dns machines (g1,g2,g3,g4). that all seems to be working. - backend files system on g[1-4] is xfs. ie, g1:/var/log/glusterfs # xfs_info /mnt/glus1 meta-data=/dev/sdb1 isize=256agcount=7, agsize=268435200 blks = sectsz=512 attr=2 data = bsize=4096 blocks=1627196928, imaxpct=5 = sunit=256swidth=2560 blks naming =version 2 bsize=4096 ascii-ci=0 log =internal bsize=4096 blocks=32768, version=2 = sectsz=512 sunit=8 blks, lazy-count=0 realtime =none extsz=4096 blocks=0, rtextents=0 - sometimes root can stat/read the file in question while the user cannot! i can remount the same NFS share to another mount point - and i can then see that with the same user. I think that may be occurring because NFS+LDAP requires a slightly different authentication scheme as compared to a NFS only setup. Please try the same test without LDAP in the middle. - sample output of g1 nfs.log file: [2011-02-18 15:27:07.201433] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/conan/hda/published/shot/backup/.svn/tmp/entries [2011-02-18 15:27:07.201445] I [io-stats.c:353:io_stats_dump_fd] glustervol1: BytesWritten : 1414 bytes [2011-02-18 15:27:07.201455] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 001024b+ : 1 [2011-02-18 15:27:07.205999] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-02-18 15:27:07.206032] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/conan/hda/published/shot/backup/.svn/props/tempfile.tmp [2011-02-18 15:27:07.210799] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-02-18 15:27:07.210824] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/conan/hda/published/shot/backup/.svn/tmp/log [2011-02-18 15:27:07.211904] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-02-18 15:27:07.211928] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /prod_data/xmas/lgl/pic/mr_all_PBR_HIGHNO_DF/035/1920x1080/mr_all_PBR_HIGHNO_DF.6084.exr [2011-02-18 15:27:07.211940] I [io-stats.c:343:io_stats_dump_fd] glustervol1: Lifetime : 8731secs, 610796usecs [2011-02-18 15:27:07.211951] I [io-stats.c:353:io_stats_dump_fd] glustervol1: BytesWritten : 2321370 bytes [2011-02-18 15:27:07.211962] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 000512b+ : 1 [2011-02-18 15:27:07.211972] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 002048b+ : 1 [2011-02-18 15:27:07.211983] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 004096b+ : 4 [2011-02-18 15:27:07.212009] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 008192b+ : 4 [2011-02-18 15:27:07.212019] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 016384b+ : 20 [2011-02-18 15:27:07.212030] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 032768b+ : 54 [2011-02-18 15:27:07.228051] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd
Re: [Gluster-users] Fwd: files not syncing up with glusterfs 3.1.2
David Lloyd wrote: I'm working with Paul on this. We did take advice on XFS beforehand, and were given the impression that it would just be a performance issue rather than things not actually working. We've got quite fast hardware, and are more comfortable with XFS that ext4 from our own experience so we did our own tests and were happy with XFS performance. Likewise, we're aware of the very poor performance of gluster with small files. We serve a lot of large files, and we're now moved most of the small files off to a normal nfs server. Again small files aren't known to break gluster are they? No. Gluster will work fine but the usual caveats about small file performance over a network file system apply. -Shehjar David On 21 February 2011 14:42, Fabricio Cannini fcann...@gmail.com wrote: Em Sexta-feira 18 Fevereiro 2011, às 23:24:10, paul simpson escreveu: hello all, i have been testing gluster as a central file server for a small animation studio/post production company. my initial experiments were using the fuse glusterfs protocol - but that ran extremely slowly for home dirs and general file sharing. we have since switched to using NFS over glusterfs. NFS has certainly seemed more responsive re. stat and dir traversal. however, i'm now being plagued with three different types of errors: 1/ Stale NFS file handle 2/ input/output errors 3/ and a new one: $ l -l /n/auto/gv1/production/conan/hda/published/OLD/ ls: cannot access /n/auto/gv1/production/conan/hda/published/OLD/shot: Remote I/O error total 0 d? ? ? ? ?? shot ...so it's a bit all over the place. i've tried rebooting both servers and clients. these issues are very erratic - they come and go. some information on my setup: glusterfs 3.1.2 g1:~ # gluster volume info Volume Name: glustervol1 Type: Distributed-Replicate Status: Started Number of Bricks: 4 x 2 = 8 Transport-type: tcp Bricks: Brick1: g1:/mnt/glus1 Brick2: g2:/mnt/glus1 Brick3: g3:/mnt/glus1 Brick4: g4:/mnt/glus1 Brick5: g1:/mnt/glus2 Brick6: g2:/mnt/glus2 Brick7: g3:/mnt/glus2 Brick8: g4:/mnt/glus2 Options Reconfigured: performance.write-behind-window-size: 1mb performance.cache-size: 1gb performance.stat-prefetch: 1 network.ping-timeout: 20 diagnostics.latency-measurement: off diagnostics.dump-fd-stats: on that is 4 servers - serving ~30 clients - 95% linux, 5% mac. all NFS. other points: - i'm automounting using NFS via autofs (with ldap). ie: gus:/glustervol1 on /n/auto/gv1 type nfs (rw,vers=3,rsize=32768,wsize=32768,intr,sloppy,addr=10.0.0.13) gus is pointing to rr dns machines (g1,g2,g3,g4). that all seems to be working. - backend files system on g[1-4] is xfs. ie, g1:/var/log/glusterfs # xfs_info /mnt/glus1 meta-data=/dev/sdb1 isize=256agcount=7, agsize=268435200 blks = sectsz=512 attr=2 data = bsize=4096 blocks=1627196928, imaxpct=5 = sunit=256swidth=2560 blks naming =version 2 bsize=4096 ascii-ci=0 log =internal bsize=4096 blocks=32768, version=2 = sectsz=512 sunit=8 blks, lazy-count=0 realtime =none extsz=4096 blocks=0, rtextents=0 - sometimes root can stat/read the file in question while the user cannot! i can remount the same NFS share to another mount point - and i can then see that with the same user. - sample output of g1 nfs.log file: [2011-02-18 15:27:07.201433] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/conan/hda/published/shot/backup/.svn/tmp/entries [2011-02-18 15:27:07.201445] I [io-stats.c:353:io_stats_dump_fd] glustervol1: BytesWritten : 1414 bytes [2011-02-18 15:27:07.201455] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 001024b+ : 1 [2011-02-18 15:27:07.205999] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-02-18 15:27:07.206032] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/conan/hda/published/shot/backup/.svn/props/tempfile.tmp [2011-02-18 15:27:07.210799] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-02-18 15:27:07.210824] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/conan/hda/published/shot/backup/.svn/tmp/log [2011-02-18 15:27:07.211904] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-02-18 15:27:07.211928] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /prod_data/xmas/lgl/pic/mr_all_PBR_HIGHNO_DF/035/1920x1080/mr_all_PBR_HIGHN O_DF.6084.exr [2011-02-18 15:27:07.211940] I [io-stats.c:343:io_stats_dump_fd] glustervol1: Lifetime : 8731secs, 610796usecs [2011-02-18 15:27:07.211951] I [io-stats.c:353:io_stats_dump_fd] glustervol1: BytesWritten : 2321370 bytes [2011-02-18 15:27:07.211962] I [io-stats.c:365:io_stats_dump_fd] glustervol1: Write 000512b+ : 1
Re: [Gluster-users] Mac / NFS problems
David Lloyd wrote: Hello, Were having issues with macs writing to our gluster system. Gluster vol info at end. On a mac, if I make a file in the shell I get the following message: smoke:hunter david$ echo hello test -bash: test: Operation not permitted I can help if you can send the nfs.log file from the /etc/glusterd directory on the nfs server. Before your mount command, set the log-level to trace for nfs server and then run the echo command above. Unmount as soon as you see the error above and email me the nfs.log. -Shehjar And the file is made but is zero size. smoke:hunter david$ ls -l test -rw-r--r-- 1 david realise 0 Mar 3 08:44 test glusterfs/nfslog logs thus: [2011-03-03 08:44:10.379188] I [io-stats.c:333:io_stats_dump_fd] glustervol1: --- fd stats --- [2011-03-03 08:44:10.379222] I [io-stats.c:338:io_stats_dump_fd] glustervol1: Filename : /production/hunter/test Then try to open the file: smoke:hunter david$ cat test and get the following messages in the log: [2011-03-03 08:51:13.957319] I [afr-common.c:716:afr_lookup_done] glustervol1-replicate-0: background meta-data self-heal triggered. path: /production/hunter/test [2011-03-03 08:51:13.959466] I [afr-self-heal-common.c:1526:afr_self_heal_completion_cbk] glustervol1-replicate-0: background meta-data self-heal completed on /production/hunter/test If I do the same test on a linux machine (nfs) it's fine. We get the same issue on all the macs. They are 10.6.6. Gluster volume is mounted: /n/auto/gv1 -rw,hard,tcp,rsize=32768,wsize=32768,intr gus:/glustervol1 Other nfs mounts on mac (from linux servers) are OK We're using LDAP to authenticate on the macs, the gluster servers aren't bound into the LDAP domain. Any ideas? Thanks David g3:/var/log/glusterfs # gluster volume info Volume Name: glustervol1 Type: Distributed-Replicate Status: Started Number of Bricks: 4 x 2 = 8 Transport-type: tcp Bricks: Brick1: g1:/mnt/glus1 Brick2: g2:/mnt/glus1 Brick3: g3:/mnt/glus1 Brick4: g4:/mnt/glus1 Brick5: g1:/mnt/glus2 Brick6: g2:/mnt/glus2 Brick7: g3:/mnt/glus2 Brick8: g4:/mnt/glus2 Options Reconfigured: performance.stat-prefetch: 1 performance.cache-size: 1gb performance.write-behind-window-size: 1mb network.ping-timeout: 20 diagnostics.latency-measurement: off diagnostics.dump-fd-stats: on ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Gluster for Oracle Data repository
Are you using the Direct nfs client in Oracle? ???(yongjoon kong)/Cloud Computing wrote: Hi all, I’m trying to use Gluster as Oracle Data home via Gluster NFS I tried it in a small size. -Made small Gluster using 2 node and using AFR -and mount from client using NFS -create tablespace and insert data into the Oracle -while doing that, I shutdown one of bricks. But it was OK anyway.( I mean there’s no error in inserting procedure) So I’m going to go on a big scale… Do you guys have any recommendation on that? ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] 3.1.3 NFS - cant overwrite certain files..
And, how do you know over-writing is failing? better still, please post the sequence of commands that resulted in this failure. Thanks Pranith Kumar. Karampuri wrote: Could you post the output of ls -l of the files for which the write op fails from the backends. Knowing the strace output of the write operation thats failing would help. Pranith. - Original Message - From: paul simpson p...@realisestudio.com To: gluster-users@gluster.org Sent: Tuesday, March 22, 2011 5:18:21 AM Subject: [Gluster-users] 3.1.3 NFS - cant overwrite certain files.. hi, i'm running 3.1.3. i'm finding that certain machines cant overwrite certain files - getting a operation not permitted. the files are owned by the same user. nothing appears on the gluster nfs log. config: g1:/var/log/glusterfs # gluster volume info Volume Name: glustervol1 Type: Distributed-Replicate Status: Started Number of Bricks: 4 x 2 = 8 Transport-type: tcp Bricks: Brick1: g1:/mnt/glus1 Brick2: g2:/mnt/glus1 Brick3: g3:/mnt/glus1 Brick4: g4:/mnt/glus1 Brick5: g1:/mnt/glus2 Brick6: g2:/mnt/glus2 Brick7: g3:/mnt/glus2 Brick8: g4:/mnt/glus2 Options Reconfigured: performance.write-behind-window-size: 1mb performance.cache-size: 1gb performance.stat-prefetch: 1 network.ping-timeout: 20 diagnostics.latency-measurement: off diagnostics.dump-fd-stats: on when mounted with FUSE - files can be overwritten. this is very disruptive - as it's stalling/breaking complex grid jobs. has anyone seen this behaviour at all? ...any ideas??? regards to all, paul ps - this also happened with 3.1.2. i just upgraded hoping that it would be fixed. ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] gluster 3.1.3 mount using nfs
Try port=38467. Let us know if it works. We'll put it up on the FAQ. Thanks. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Monday, March 28, 2011 5:54:06 PM Subject: [Gluster-users] gluster 3.1.3 mount using nfs Hi all, I setup the gluster filesystem and I want to mount the gluster volume using nfs in unix system. My machine is hp-ux (11.23) I put command like below but it has error test14:/mount -F nfs -o proto=tcp,port=38465,vers=3,llock 150.2.226.26:/temp /mnt nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs mount: retry: retrying(1) for: /mnt after 5 seconds nfs mount: retry: giving up on: /mnt and it was ok when I tried it in linux machine. I attached the volume setting Plz help me out. Volume Name: isi Type: Distributed-Replicate Status: Started Number of Bricks: 2 x 2 = 4 Transport-type: tcp Bricks: Brick1: node003:/data02 Brick2: node004:/data02 Brick3: node005:/data02 Brick4: node006:/data02 Options Reconfigured: nfs.rpc-auth-allow: 150.2.223.249 nfs.rpc-auth-unix: on Cloud Computing Business Team Andrew Kong Assistant Manager | andrew.k...@sk.com| T:+82-2-6400-4328 | M:+82-10-8776-5025 SK u-Tower, 25-1, Jeongja-dong, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-844, Korea SK CChttp://www.skcc.co.kr/ : Abouthttp://www.skcc.co.kr/user/common/userContentViewer.vw?menuID=KRCA0300 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] gluster 3.1.3 mount using nfs
On the HP-UX machine, whats the output of: showmount -e 150.2.226.26 - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Tuesday, March 29, 2011 5:31:28 AM Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs It doesn't work. For my guess, it's probably connected with multiple networks. By gluster bricks is on 192.168.x.x one of the brick server has public network like 150.2.x.x And the unix server's ip is 150.2.x.236 and try to mount the brick via nfs through that network Line Is it possible configurations? -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, March 29, 2011 1:42 AM To: 공용준(yongjoon kong)/Cloud Computing 사업담당 Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs Try port=38467. Let us know if it works. We'll put it up on the FAQ. Thanks. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Monday, March 28, 2011 5:54:06 PM Subject: [Gluster-users] gluster 3.1.3 mount using nfs Hi all, I setup the gluster filesystem and I want to mount the gluster volume using nfs in unix system. My machine is hp-ux (11.23) I put command like below but it has error test14:/mount -F nfs -o proto=tcp,port=38465,vers=3,llock 150.2.226.26:/temp /mnt nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs mount: retry: retrying(1) for: /mnt after 5 seconds nfs mount: retry: giving up on: /mnt and it was ok when I tried it in linux machine. I attached the volume setting Plz help me out. Volume Name: isi Type: Distributed-Replicate Status: Started Number of Bricks: 2 x 2 = 4 Transport-type: tcp Bricks: Brick1: node003:/data02 Brick2: node004:/data02 Brick3: node005:/data02 Brick4: node006:/data02 Options Reconfigured: nfs.rpc-auth-allow: 150.2.223.249 nfs.rpc-auth-unix: on Cloud Computing Business Team Andrew Kong Assistant Manager | andrew.k...@sk.com| T:+82-2-6400-4328 | M:+82-10-8776-5025 SK u-Tower, 25-1, Jeongja-dong, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-844, Korea SK CChttp://www.skcc.co.kr/ : Abouthttp://www.skcc.co.kr/user/common/userContentViewer.vw?menuID=KRCA0300 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] gluster 3.1.3 mount using nfs
OK. See if the following steps help on the HP-UX client: 1. Add the line NFS_TCP=1 in file /etc/rc.config.d/nfsconf 2. /sbin/init.d/nfs.client stop; 3. /sbin/init.d/nfs.client start 4. Now run the mount command, first without any options and if that does not work, try with the options including port=38467. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: Shehjar Tikoo shehj...@gluster.com Cc: gluster-users@gluster.org Sent: Tuesday, March 29, 2011 1:39:23 PM Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs I got this result showmount -e 150.2.226.26 export list for 150.2.226.26: /vbs * /vms * And /vms , /vbs cannot be mounted. -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, March 29, 2011 4:08 PM To: 공용준(yongjoon kong)/Cloud Computing 사업담당 Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs On the HP-UX machine, whats the output of: showmount -e 150.2.226.26 - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Tuesday, March 29, 2011 5:31:28 AM Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs It doesn't work. For my guess, it's probably connected with multiple networks. By gluster bricks is on 192.168.x.x one of the brick server has public network like 150.2.x.x And the unix server's ip is 150.2.x.236 and try to mount the brick via nfs through that network Line Is it possible configurations? -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, March 29, 2011 1:42 AM To: 공용준(yongjoon kong)/Cloud Computing 사업담당 Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs Try port=38467. Let us know if it works. We'll put it up on the FAQ. Thanks. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Monday, March 28, 2011 5:54:06 PM Subject: [Gluster-users] gluster 3.1.3 mount using nfs Hi all, I setup the gluster filesystem and I want to mount the gluster volume using nfs in unix system. My machine is hp-ux (11.23) I put command like below but it has error test14:/mount -F nfs -o proto=tcp,port=38465,vers=3,llock 150.2.226.26:/temp /mnt nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs mount: retry: retrying(1) for: /mnt after 5 seconds nfs mount: retry: giving up on: /mnt and it was ok when I tried it in linux machine. I attached the volume setting Plz help me out. Volume Name: isi Type: Distributed-Replicate Status: Started Number of Bricks: 2 x 2 = 4 Transport-type: tcp Bricks: Brick1: node003:/data02 Brick2: node004:/data02 Brick3: node005:/data02 Brick4: node006:/data02 Options Reconfigured: nfs.rpc-auth-allow: 150.2.223.249 nfs.rpc-auth-unix: on Cloud Computing Business Team Andrew Kong Assistant Manager | andrew.k...@sk.com| T:+82-2-6400-4328 | M:+82-10-8776-5025 SK u-Tower, 25-1, Jeongja-dong, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-844, Korea SK CChttp://www.skcc.co.kr/ : Abouthttp://www.skcc.co.kr/user/common/userContentViewer.vw?menuID=KRCA0300 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] 회신: gluster 3.1.3 mount using nfs
- Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: Shehjar Tikoo shehj...@gluster.com Cc: gluster-users@gluster.org Sent: Tuesday, March 29, 2011 4:28:33 PM Subject: [Gluster-users] 회신: gluster 3.1.3 mount using nfs yes, you were right. but after i shutdown the nfs server on 150.2.226.21 You have to restart Gluster NFS server after shutting down the kernel's nfs server. -Shehjar i still has error message like below test14:/mount -F nfs 150.2.226.21:/vms /mnt nfs mount: get_fh: 150.2.226.21:: RPC: Program not registered nfs mount: get_fh: 150.2.226.21:: RPC: Program not registered nfs mount: retry: retrying(1) for: /mnt after 5 seconds nfs mount: retry: giving up on: /mnt can you tell me that what's different from linux and unix nfs protocol ? cause, there's no problem with linux 보낸 사람: Shehjar Tikoo [shehj...@gluster.com] 보낸 날짜: 2011년 3월 29일 화요일 오후 6:31 받는 사람: 공용준(yongjoon kong)/Cloud Computing 사업담당 참조: gluster-users@gluster.org 제목: Re: [Gluster-users] gluster 3.1.3 mount using nfs It looks like you have another nfs server running on 150.2.226.21 because there is no mountd that is part of Gluster NFS server. Please shutdown this mountd processes and then try mounting again. If that doesnt work, try mounting with the options you had used earlier. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: Shehjar Tikoo shehj...@gluster.com Cc: gluster-users@gluster.org Sent: Tuesday, March 29, 2011 2:46:31 PM Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs The message is changed like below test14:/mount -F nfs 150.2.226.21:/vms /mnt Permission denied And from the 150.2.226.21 says Mar 29 18:17:23 skt-cldpap001 mountd[22080]: refused mount request from 150.2.223.249 for /vms (/): not exported -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, March 29, 2011 5:45 PM To: 공용준(yongjoon kong)/Cloud Computing 사업담당 Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs OK. See if the following steps help on the HP-UX client: 1. Add the line NFS_TCP=1 in file /etc/rc.config.d/nfsconf 2. /sbin/init.d/nfs.client stop; 3. /sbin/init.d/nfs.client start 4. Now run the mount command, first without any options and if that does not work, try with the options including port=38467. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: Shehjar Tikoo shehj...@gluster.com Cc: gluster-users@gluster.org Sent: Tuesday, March 29, 2011 1:39:23 PM Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs I got this result showmount -e 150.2.226.26 export list for 150.2.226.26: /vbs * /vms * And /vms , /vbs cannot be mounted. -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, March 29, 2011 4:08 PM To: 공용준(yongjoon kong)/Cloud Computing 사업담당 Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs On the HP-UX machine, whats the output of: showmount -e 150.2.226.26 - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Tuesday, March 29, 2011 5:31:28 AM Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs It doesn't work. For my guess, it's probably connected with multiple networks. By gluster bricks is on 192.168.x.x one of the brick server has public network like 150.2.x.x And the unix server's ip is 150.2.x.236 and try to mount the brick via nfs through that network Line Is it possible configurations? -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Tuesday, March 29, 2011 1:42 AM To: 공용준(yongjoon kong)/Cloud Computing 사업담당 Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] gluster 3.1.3 mount using nfs Try port=38467. Let us know if it works. We'll put it up on the FAQ. Thanks. - Original Message - From: 공용준(yongjoon kong)/Cloud Computing 사업담당 andrew.k...@sk.com To: gluster-users@gluster.org gluster-users@gluster.org Sent: Monday, March 28, 2011 5:54:06 PM Subject: [Gluster-users] gluster 3.1.3 mount using nfs Hi all, I setup the gluster filesystem and I want to mount the gluster volume using nfs in unix system. My machine is hp-ux (11.23) I put command like below but it has error test14:/mount -F nfs -o proto=tcp,port=38465,vers=3,llock 150.2.226.26:/temp /mnt nfs mount: get_fh: 150.2.226.26:: RPC: Program not registered nfs
Re: [Gluster-users] Disabling NFS
Mike Hanby wrote: Strange, I do have the fuse and gluster-fuse / gluster-core packages installed on the client. I can mount the volume using the gluster nas-srv-01:/users /users glusterfs defaults,_netdev 0 0 Maybe I just need to figure out how to configure builtin Gluster NFS to export like I want. See http://gluster.com/community/documentation/index.php/Gluster_3.1_NFS_Guide I'm trying to ensure that from the Gluster servers, I have fine control over which ip addresses can mount specific parts of the volume, similar to what can be done via /etc/exports See rpc-auth.addr option in the NFS options section in the 3.1.3 release notes: http://download.gluster.com/pub/gluster/glusterfs/3.1/3.1.3/Gluster_FS_Release_Notes_3.1.3.pdf -Shehjar ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] How do I diagnose what's going wrong with a Gluster NFS mount?
If you can file a bug, we'll take it from there. Thanks. Whit Blauvelt wrote: Hi, Has anyone even seen this before - an NFS mount through Gluster that gets the filesystem size wrong and is otherwise garbled and dangerous? Is there a way within Gluster to fix it, or is the lesson that Gluster's NFS sometimes can't be relied on? What have the experiences been running an external NFS daemon with Gluster? Is that fairly straightforward? Might like to get the advantages of NFS4 anyhow. Thanks, Whit - Forwarded message from Whit Blauvelt whit.glus...@transpect.com - Date: Sat, 28 May 2011 22:33:55 -0400 From: Whit Blauvelt whit.glus...@transpect.com To: gluster-users@gluster.org Subject: nfs mount in error, wrong filesystem size shown User-Agent: Mutt/1.5.21 (2010-09-15) I've got a couple of servers with several mirrored gluster volumes. Two work fine from all perspectives. One, the most recently set up, mounts remotely as glusterfs, but fails badly as nfs. The mount appears to work when requested, but the filesystem size shown in totally wrong and it is not in fact accessible. This is with 3.1.4: So we have for instance on one external system: 192.168.1.242:/std 309637120 138276672 155631808 48% /mnt/std 192.168.1.242:/store 19380692 2860644 15543300 16% /mnt/store where the first nfs mount is correct and working, but the second is way off. That was the same result as when /store was nfs mounted to another system too. But on that same other system, /store mounts correctly as glusterfs: vm2:/store536704000 14459648 494981376 3% /mnt/store with the real size shown, and the filesystem fully accessible. The erroneous mount is also apparently dangerous. I tried writing a file to it to see what would happen, and it garbaged the underlying filesystems. So I did a full reformatting and recreation of the gluster volume before retrying at that point - and still got the bad nfs mount for it. The bad nfs mount happens no matter which of the two servers in the gluster cluster the mount uses, too. Any ideas what I'm hitting here? For the present purpose, we need to be able to mount nfs, as we need some Macs to mount it. Thanks, Whit - End forwarded message - ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] Glusterfs 3.2.0 NFS Problem
This can happen if all your servers were unreachable for a few seconds. The situation must have rectified during the restart. We could confirm if you change the log level on nfs to DEBUG and send us the log. Thanks -Shehjar Jurgen Winkler wrote: Hi, i noticed a strange behavior with NFS and Glusterfs 3.2.0 , 3 of our Servers are loosing the Mount but when you restart the Volume on the Server it works again without a remount. On the server i noticed this entries in the Glusterfs/Nfs log-file when the mount on the Client becomes unavailable : [2011-06-08 14:37:02.568693] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:02.569212] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:02.611910] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:02.624477] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.288272] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.296150] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.309247] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.320939] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.321786] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.333609] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.334089] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.344662] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.352666] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.354195] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.360446] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.369331] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.471556] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.480013] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:05.639700] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:05.652535] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.578469] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.588949] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.590395] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.591414] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.591932] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.592596] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.639317] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:07.652919] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.332435] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.340622] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.349360] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.349550] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.360445] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.369497] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.369752] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.382097] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:09.382387] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up Thx for the help ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS problem
(GlusterFS 3.1) op(STAT(1)) called at 2011-06-09 17:01:35.784870 [2011-06-09 17:01:35.817918] I [client3_1-fops.c:411:client3_1_stat_cbk] 0-poolsave-client-1: remote operation failed: Transport endpoint is not connected [2011-06-09 17:01:35.817969] I [client.c:1883:client_rpc_notify] 0-poolsave-client-1: disconnected [2011-06-09 17:01:35.817988] E [afr-common.c:2546:afr_notify] 0-poolsave-replicate-0: All subvolumes are down. Going offline until atleast one of them comes back up. [2011-06-09 17:01:35.818007] E [socket.c:1685:socket_connect_finish] 0-poolsave-client-1: connection to 10.68.217.86:24011 failed (Connection refused) [2011-06-09 17:01:35.818606] I [afr.h:838:AFR_LOCAL_INIT] 0-poolsave-replicate-0: no subvolumes up [2011-06-09 17:01:35.819129] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /uvs00/log: no child is up [2011-06-09 17:01:35.819354] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /uvs00/log: no child is up [2011-06-09 17:01:35.820090] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /uvs00: no child is up [2011-06-09 17:01:35.820760] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.821212] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.821600] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.822123] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.822511] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.822975] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.823286] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.823583] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:35.823857] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:47.518006] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up [2011-06-09 17:01:49.39204] E [socket.c:1685:socket_connect_finish] 0-poolsave-client-0: connection to 10.68.217.85:24014 failed (Connection refused) [2011-06-09 17:01:49.136932] I [afr-inode-read.c:270:afr_stat] 0-poolsave-replicate-0: /: no child is up Message: 7 Date: Thu, 9 Jun 2011 12:56:39 +0530 From: Shehjar Tikoo shehj...@gluster.com Subject: Re: [Gluster-users] Glusterfs 3.2.0 NFS Problem To: J?rgen Winkler juergen.wink...@xidras.com Cc: gluster-users@gluster.org Message-ID: 4df075af.3040...@gluster.com Content-Type: text/plain; charset=us-ascii; format=flowed This can happen if all your servers were unreachable for a few seconds. The situation must have rectified during the restart. We could confirm if you change the log level on nfs to DEBUG and send us the log. Thanks -Shehjar Jurgen Winkler wrote: Hi, i noticed a strange behavior with NFS and Glusterfs 3.2.0 , 3 of our Servers are loosing the Mount but when you restart the Volume on the Server it works again without a remount. On the server i noticed this entries in the Glusterfs/Nfs log-file when the mount on the Client becomes unavailable : [2011-06-08 14:37:02.568693] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:02.569212] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:02.611910] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:02.624477] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.288272] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.296150] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.309247] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.320939] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.321786] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.333609] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.334089] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.344662] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.352666] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.354195] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.360446] I [afr-inode-read.c:270:afr_stat] 0-ksc-replicate-0: /: no child is up [2011-06-08 14:37:04.369331] I [afr-inode-read.c:270:afr_stat
Re: [Gluster-users] Failed to map FH to vol
They have no meaning unless you get a Stale NFs file handle error on a *NFS* mount. It generally means that the NFS client is trying to access a volume/share that does not exist anymore. I've been thinking of reducing the log-level of this message for sometime. Do you think it is a problem enough to change the log-level in the source? -Shehjar Daniel Manser wrote: Dear list, What do these messages mean? My client can't write and these messages are filling up the nfs.log: [2011-06-14 13:38:42.826673] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:38:51.826871] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:00.827037] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:09.827248] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:18.827401] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:27.827578] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:36.827783] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:45.829407] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:39:54.828145] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol [2011-06-14 13:40:03.828344] E [nfs3.c:724:nfs3_getattr] 0-nfs-nfsv3: Failed to map FH to vol Thanks, Daniel ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS expose subfolders only
Directory exports are enabled by default. You just need to mount using /bkfarm/00 as the export dir, not /00. -Shehjar Thai. Ngo Bao wrote: Anush, thanks for the quick reply. Below is the output of showmount at server side: [root@GS_BackupFarm_Cluster01 ~]# showmount -e localhost Export list for localhost: /bkfarm/00 * Output from netstat: [root@GS_BackupFarm_Cluster01 ~]# netstat -vtlpn Active Internet connections (only servers) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp0 0 0.0.0.0:38465 0.0.0.0:* LISTEN 1531/glusterfs tcp0 0 0.0.0.0:38466 0.0.0.0:* LISTEN 1531/glusterfs tcp0 0 0.0.0.0:38467 0.0.0.0:* LISTEN 1531/glusterfs tcp0 0 0.0.0.0:805 0.0.0.0:* LISTEN 4015/rpc.statd tcp0 0 0.0.0.0:111 0.0.0.0:* LISTEN 3934/portmap tcp0 0 :::24007 :::*LISTEN 22996/glusterd tcp0 0 :::24009 :::*LISTEN 8653/glusterfsd -client side --- [root@GSO_DB_Local1 ~]# showmount -e localhost mount clntudp_create: RPC: Program not registered [root@GSO_DB_Local1 ~]# mount -o vers=3 bkf3:/00 /bkfarm/ mount: bkf3:/00 failed, reason given by server: No such file or directory Any ideas? Thanks, ~Thai *From:* Anush Shetty [mailto:an...@gluster.com] *Sent:* Monday, November 14, 2011 5:40 PM *To:* Thai. Ngo Bao; gluster-users@gluster.org *Subject:* RE: NFS expose subfolders only Hi, Please make sure that nfs-kernel-server isn't running. Can you pase your showmount -e output? The right way to mount Gluster NFS is, mount -o vers=3 bkf3:/00 /bkfarm/ - Anush *From:* Thai. Ngo Bao [tha...@vng.com.vn] *Sent:* 14 November 2011 16:07:05 *To:* Anush Shetty; gluster-users@gluster.org *Subject:* RE: NFS expose subfolders only Hi, I have tried the trick several times but had no success so far. Below is the info of my testing environment [root@GS_BackupFarm_Cluster01 ~]# gluster volume set bkfarm nfs.export-dir /00 Set volume successful [root@GS_BackupFarm_Cluster01 ~]# gluster volume set bkfarm nfs.export-volumes off Set volume successful [root@GS_BackupFarm_Cluster01 ~]# gluster volume info Volume Name: bkfarm Type: Distributed-Replicate Status: Started Number of Bricks: 4 x 2 = 8 Transport-type: tcp Bricks: Brick1: bkf1:/sfarm Brick2: bkf2:/sfarm Brick3: bkf3:/sfarm Brick4: bkf4:/sfarm Brick5: bkf5:/sfarm Brick6: bkf6:/sfarm Brick7: bkf7:/sfarm Brick8: bkf8:/sfarm Options Reconfigured: nfs.disable: Off nfs.export-dir: /00 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
Re: [Gluster-users] NFS expose subfolders only
Like I said earlier, all folders are exported by default, just that they dont show up on showmount -e output unless explicitly set using the nfs.export-dir option. This means, if you know that a folder exists on the Gluster volume, you can mount it directly like I had shown in the previous email. -Shehjar Thai. Ngo Bao wrote: Hi, Is it possible to expose multiple sub-folders using nfs.export-dir? I do not have an access to my test environment right now, so I guess something similar should be done: gluster volume set volume_name nfs.export-dir /sub-folder1 /sub-folder2 ... ? Any insight into this is much appreciated, and I will definitely figure out what should be done tomorrow for the case. Thanks, ~Thai From: gluster-users-boun...@gluster.org [gluster-users-boun...@gluster.org] On Behalf Of Thai. Ngo Bao [tha...@vng.com.vn] Sent: Monday, November 14, 2011 6:02 PM To: Shehjar Tikoo Cc: gluster-users@gluster.org Subject: Re: [Gluster-users] NFS expose subfolders only Bingo. It works. Shehjar, thanks for your hint. ~Thai -Original Message- From: Shehjar Tikoo [mailto:shehj...@gluster.com] Sent: Monday, November 14, 2011 5:55 PM To: Thai. Ngo Bao Cc: Anush Shetty; gluster-users@gluster.org Subject: Re: [Gluster-users] NFS expose subfolders only Directory exports are enabled by default. You just need to mount using /bkfarm/00 as the export dir, not /00. -Shehjar Thai. Ngo Bao wrote: Anush, thanks for the quick reply. Below is the output of showmount at server side: [root@GS_BackupFarm_Cluster01 ~]# showmount -e localhost Export list for localhost: /bkfarm/00 * Output from netstat: [root@GS_BackupFarm_Cluster01 ~]# netstat -vtlpn Active Internet connections (only servers) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp0 0 0.0.0.0:38465 0.0.0.0:* LISTEN 1531/glusterfs tcp0 0 0.0.0.0:38466 0.0.0.0:* LISTEN 1531/glusterfs tcp0 0 0.0.0.0:38467 0.0.0.0:* LISTEN 1531/glusterfs tcp0 0 0.0.0.0:805 0.0.0.0:* LISTEN 4015/rpc.statd tcp0 0 0.0.0.0:111 0.0.0.0:* LISTEN 3934/portmap tcp0 0 :::24007 :::*LISTEN 22996/glusterd tcp0 0 :::24009 :::*LISTEN 8653/glusterfsd -client side --- [root@GSO_DB_Local1 ~]# showmount -e localhost mount clntudp_create: RPC: Program not registered [root@GSO_DB_Local1 ~]# mount -o vers=3 bkf3:/00 /bkfarm/ mount: bkf3:/00 failed, reason given by server: No such file or directory Any ideas? Thanks, ~Thai *From:* Anush Shetty [mailto:an...@gluster.com] *Sent:* Monday, November 14, 2011 5:40 PM *To:* Thai. Ngo Bao; gluster-users@gluster.org *Subject:* RE: NFS expose subfolders only Hi, Please make sure that nfs-kernel-server isn't running. Can you pase your showmount -e output? The right way to mount Gluster NFS is, mount -o vers=3 bkf3:/00 /bkfarm/ - Anush *From:* Thai. Ngo Bao [tha...@vng.com.vn] *Sent:* 14 November 2011 16:07:05 *To:* Anush Shetty; gluster-users@gluster.org *Subject:* RE: NFS expose subfolders only Hi, I have tried the trick several times but had no success so far. Below is the info of my testing environment [root@GS_BackupFarm_Cluster01 ~]# gluster volume set bkfarm nfs.export-dir /00 Set volume successful [root@GS_BackupFarm_Cluster01 ~]# gluster volume set bkfarm nfs.export-volumes off Set volume successful [root@GS_BackupFarm_Cluster01 ~]# gluster volume info Volume Name: bkfarm Type: Distributed-Replicate Status: Started Number of Bricks: 4 x 2 = 8 Transport-type: tcp Bricks: Brick1: bkf1:/sfarm Brick2: bkf2:/sfarm Brick3: bkf3:/sfarm Brick4: bkf4:/sfarm Brick5: bkf5:/sfarm Brick6: bkf6:/sfarm Brick7: bkf7:/sfarm Brick8: bkf8:/sfarm Options Reconfigured: nfs.disable: Off nfs.export-dir: /00 - No virus found in this message. Checked by AVG - www.avg.com Version: 2012.0.1869 / Virus Database: 2092/4615 - Release Date: 11/13/11 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users - No virus found in this message. Checked by AVG - www.avg.com Version: 2012.0.1869 / Virus Database: 2092/4615 - Release Date: 11/13/11 ___ Gluster-users mailing list Gluster-users@gluster.org http://gluster.org/cgi-bin/mailman/listinfo/gluster-users