From sbarthel at clusterfs.com Thu Oct 4 02:38:16 2007 From: sbarthel at clusterfs.com (Sheila Barthel) Date: Wed, 3 Oct 2007 20:38:16 -0600 Subject: [Lustre-devel] Announcement for Lustre 1.6 Operations Manual (document edition v1_8) Message-ID: <837874b50710031938x1a5af63ehe0a117a465fde759@mail.gmail.com> Dear Lustre users: We are pleased to announce an updated document edition (v1_8) of the Lustre(r) 1.6 Operations Manual, is available in both PDF and HTML formats at http://www.manual.lustre.org. This edition of the Operations Manual includes the following enhancements: * New POSIX chapter * New Benchmarking chapter * New Lustre Recovery chapter * Updated content in Configuring Quotas chapter * Updated content in More Complicated Configurations chapter * Updated content in Lustre /proc chapter * Updated content in RAID chapter * Updated content in Configuring Lustre Examples chapter * Updated content in Configuration Files Module Parameters chapter * Merged MXLND information in Lustre manual If you have any questions, comments, or recommended edits to the manual, please file a ticket in Bugzilla, https://bugzilla.lustre.org/enter_bug.cgi, specifying Documentation as the product. Sun Lustre Documentation Team -------------- next part -------------- An HTML attachment was scrubbed... URL: From serge at terascala.com Tue Oct 9 20:52:12 2007 From: serge at terascala.com (serge at terascala.com) Date: Tue, 9 Oct 2007 14:52:12 -0600 Subject: [Lustre-devel] [Bug 13840] New: Problems compiling with OFED 1.2.5 Message-ID: <200710092052.l99KqCI1019435@shell.clusterfs.com> Please don't reply to lustre-devel. Instead, comment in Bugzilla by using the following link: https://bugzilla.lustre.org/show_bug.cgi?id=13840 When I try to compile the lnet module o2ib with OFED 1.2.5.1, I get the following error: /local/home/serge/2.6.16-46/bnx2/lustre/1.6.2/trunk/lustre_shadow_dir/lnet/klnds/o2iblnd/o2iblnd.c In file included from /usr/local/home/serge/2.6.16-46/bnx2/lustre/1.6.2/trunk/lustre_shadow_dir/lnet/klnds/o2iblnd/o2iblnd.c:24: /usr/local/home/serge/2.6.16-46/bnx2/lustre/1.6.2/trunk/lustre_shadow_dir/lnet/klnds/o2iblnd/o2iblnd.h:330: error: parse error before "DECLARE_PCI_UNMAP_ADDR" Serge Pirotte Terascala From ohara at rd.scei.sony.co.jp Mon Oct 15 09:54:20 2007 From: ohara at rd.scei.sony.co.jp (Kazuki Ohara) Date: Mon, 15 Oct 2007 18:54:20 +0900 Subject: [Lustre-devel] Is dump_kernel fixed? Message-ID: <471338CC.4020701@rd.scei.sony.co.jp> Hi, The status of the bug #12782 is RESOLVED, but the resolution seems to be not merged into the latest 1.6.3. If the resolution is avalilable for now, where can I get it from ? Regards, Kazuki Ohara From bjohanso at psc.edu Tue Oct 16 14:49:50 2007 From: bjohanso at psc.edu (bjohanso at psc.edu) Date: Tue, 16 Oct 2007 08:49:50 -0600 Subject: [Lustre-devel] [Bug 13607] lnet router RDMA too fragmented: 128/256 src 128/256 dst frags Message-ID: <200710161449.l9GEnoOl007101@shell.clusterfs.com> Please don't reply to lustre-devel. Instead, comment in Bugzilla by using the following link: https://bugzilla.lustre.org/show_bug.cgi?id=13607 Any news on this one? From shadow at clusterfs.com Wed Oct 17 14:50:41 2007 From: shadow at clusterfs.com (shadow at clusterfs.com) Date: Wed, 17 Oct 2007 08:50:41 -0600 Subject: [Lustre-devel] [Bug 13913] New: ASSERTION(atomic_read(&ctxt_->loc_refcount) > 0) failed Message-ID: <200710171450.l9HEofZK001376@shell.clusterfs.com> Please don't reply to lustre-devel. Instead, comment in Bugzilla by using the following link: https://bugzilla.lustre.org/show_bug.cgi?id=13913 while run sanity on loop devices i got assetion. ASSERTION(atomic_read(&ctxt_->loc_refcount) > 0) failed this after loop device error and got ENOMEM. From jmferrer at caton.es Mon Oct 22 10:56:37 2007 From: jmferrer at caton.es (jmferrer) Date: Mon, 22 Oct 2007 12:56:37 +0200 Subject: [Lustre-devel] The mds_connect operation failed with -11 Message-ID: <471C81E5.5000601@caton.es> Hi, list: I'm trying configure lustre with: 1 MGS -------------> 192.168.3.100 with mkfs.lustre --mgs /dev/md1 ; mount -t lustre ... 1 MDT ------------> 192.168.3.101 with mkfs.lustre --fsname=datafs00 --mdt --mgsnode=192.168.3.100 /dev/sda3 ; mount -t lustre ... 4 ost -----------> 192.168.3.102-104 with mkfs.lustre --fsname=datafs00 --ost --mgsnode=192.168.3.100 at tcp0 /dev/sda3 ; mount -t lustre..... foreach node But when I try mount from any node: LOG IN NODE: LustreError: 4743:0:(obd_mount.c:1927:lustre_fill_super()) Unable to mount (-22) LustreError: 11-0: an error occurred while communicating with 192.168.3.101 at tcp. The mds_connect operation failed with -11 LustreError: Skipped 1 previous similar message LustreError: 11-0: an error occurred while communicating with 192.168.3.101 at tcp. The mds_connect operation failed with -11 LustreError: Skipped 5 previous similar messages LustreError: 11-0: an error occurred while communicating with 192.168.3.101 at tcp. The mds_connect operation failed with -11 LustreError: Skipped 10 previous similar messages LustreError: 11-0: an error occurred while communicating with 192.168.3.101 at tcp. The mds_connect operation failed with -11 LustreError: Skipped 20 previous similar messages LOG IN MGS: --------- thanks and sorry for my english. From Jingjing.Li at Sun.COM Fri Oct 26 07:53:31 2007 From: Jingjing.Li at Sun.COM (Karen Li) Date: Fri, 26 Oct 2007 15:53:31 +0800 Subject: [Lustre-devel] Last Chance! Lustre Level I training sessions on November 7-9 in Reno Message-ID: <47219CFB.30102@sun.com> Lustre Administration & Support training will be offered on November 7-9 in Reno, Nevada. Only a few spaces remaining! Click here to register Today! http://www.acteva.com/go/clusterfilesystems Lustre Administration & Support (CFS1A) This three day course is ideal for people new to the Lustre File System and will cover a host of introductory architectural, administration and support issues. The session will provide a hands-on experience and culminate with a Level I certification exam. Course Content: Day 1: * Understanding Lustre * Architecting and profiling a Lustre Cluster Day 2: * Configuring Lustre * Hands on Lab Day 3: * Troubleshooting Lustre * CFS Lustre Level I certification exam Price: US$3,000 for 3 days of in-depth technical hands-on instruction Location * Circus Circus Hotel * 500 N. Sierra Street, Reno, NV 89503 * Phone: 1-775-329-0711 * Toll-free: 1-800-648-5010 * http://www.circusreno.com/ ENROLL NOW! http://www.acteva.com/go/clusterfilesystems Questions regarding Lustre courses can be sent to training at clusterfs.com Visit www.clusterfs.com/events.html for more information. From dsternkopf at hpce.nec.com Fri Oct 26 09:24:34 2007 From: dsternkopf at hpce.nec.com (Danny Sternkopf) Date: Fri, 26 Oct 2007 11:24:34 +0200 Subject: [Lustre-devel] Lustre - HA: lc_hb bug Message-ID: <4721B252.1010808@hpce.nec.com> If lustre_config configures more than one failover pair the haresource will be wrong except for the first failover pair. lc_hb generates the haresource on the node running lustre_config in /tmp/heartbeat/haresources.lustre. Later it will be copied to haresources.lustre.. For the 2nd failover pair the haresource configuration will be added to haresources.lustre which is wrong. This file should be cleared or removed before. The following quick fix seems to help in line 436: [root at dmds1 config]# sdiff -s /usr/sbin/lc_hb /usr/sbin/lc_hb.orig # Clean /tmp/heartbeat/haresources.lustre first < rm -f ${HARES_LUSTRE} < I'am not 100% sure if it is okay for all cases. Best regards, Danny From Jingjing.Li at Sun.COM Tue Oct 30 05:43:56 2007 From: Jingjing.Li at Sun.COM (Karen Li) Date: Tue, 30 Oct 2007 13:43:56 +0800 Subject: [Lustre-devel] Last Chance! Lustre Level I training sessions on November 7-9 in Reno Message-ID: <4726C49C.60800@sun.com> Lustre Administration & Support training will be offered on November 7-9 in Reno, Nevada. Only a few spaces remaining! Click here to register today! http://www.acteva.com/go/clusterfilesystems Lustre Administration & Support (CFS1A) This three day course is ideal for people new to the Lustre File System and will cover a host of introductory architectural, administration and support issues. The session will provide a hands-on experience and culminate with a Level I certification exam. Course Content: Day 1: * Understanding Lustre * Architecting and profiling a Lustre Cluster Day 2: * Configuring Lustre * Hands on Lab Day 3: * Troubleshooting Lustre * CFS Lustre Level I certification exam Price: US$3,000 for 3 days of in-depth technical hands-on instruction Location * Circus Circus Hotel * 500 N. Sierra Street, Reno, NV 89503 * Phone: 1-775-329-0711 * Toll-free: 1-800-648-5010 * http://www.circusreno.com/ ENROLL NOW! http://www.acteva.com/go/clusterfilesystems Questions regarding Lustre courses can be sent to training at clusterfs.com Visit www.clusterfs.com/events.html for more information.