Re: [Lustre-discuss] Quota question
On Wed, 4 Jun 2008 12:42:47 +0200 Enrico Morelli [EMAIL PROTECTED] wrote: On Wed, 04 Jun 2008 14:11:47 +0400 Andrew Perepechko [EMAIL PROTECTED] wrote: Enrico, what version of Lustre are you using? 1.6.4.3 Andrew. On Wednesday 04 June 2008 12:51:26 Enrico Morelli wrote: Dear all, I've a problem that I don't understand. Disk quotas for user hsco1 (uid 1250): Filesystem kbytes quota limit grace files quota limit grace /lustre_homes/ 13204576 3000 3500 60309 0 0 spfs-MDT_UUID 14760 0 102400 60309 0 0 spfs-OST_UUID 3648336 0 3788800 spfs-OST0001_UUID 2050408 0 2150400 spfs-OST0002_UUID 3660516 0 3788800 spfs-OST0003_UUID 3830556* 0 1 For me the user is under the quota (13GB data vs 30GB of quota), but for the system is out of quota. Where is the problem? Thanks I don't understand the last row displayed by lfs quota: spfs-OST0003_UUID 3830556* 0 1 This is my df: /dev/data_se/lustre_mgs_mdt 45871740 1515804 41734496 4% /lustre_mgs_mdt /dev/data_se/data_local_a 516061624 237605788 252241436 49% /data_local_a /dev/data_se/data_local_b 516061624 228373408 261473816 47% /data_local_b /dev/data_se/data_local_c 516061624 283749608 206097616 58% /data_local_c /dev/data_se/data_local_d 278673480 167750860 96766844 64% /data_local_d [EMAIL PROTECTED]:/spfs 1826858352 917479664 816579452 53% /lustre_homes The manual say: Note - Values appended with “*” show the limit that has been over-used (exceeding the quota), and receives this message Disk quota exceeded. How can I solve this problem? --- --- (o_ (o_//\ Coltivate Linux che tanto Windows si pianta da solo. (/)_ V_/_ +--+ | ENRICO MORELLI | email: [EMAIL PROTECTED] | | * * * *| phone: +39 055 4574269 | | University of Florence| fax : +39 055 4574253 | | CERM - via Sacconi, 6 - 50019 Sesto Fiorentino (FI) - ITALY| +--+ ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] ldiskfs kernel bug in ldiskfs_mb_use_best_found
https://bugzilla.lustre.org/show_bug.cgi?id=15932 thanks, Alex Aaron Knister wrote: For my own curiosity's sake, is there a bugzilla id associated with it? I'm curious to know what the problem is/was. On Jun 5, 2008, at 8:50 AM, Alex Zhuravlev wrote: we've been investigating one issue which can be a source of your problem. fortunately we can reproduce that. thanks, Alex Aaron Knister wrote: Unfortunately the monkeys maintaining this setup (I left the company last week) have managed to destroy one of the 13 terabyte arrays...so they've got bigger issues right now. Thanks for your help though. On Jun 5, 2008, at 3:18 AM, Alex Zhuravlev wrote: hmm. could you provide me with ldiskfs.ok then? thanks, Alex Aaron Knister wrote: I'm not sure how to obtain that. I'm running whatever was distributed with 1.6.4.3 for rhel5-x86_64. On Jun 2, 2008, at 8:08 AM, Alex Zhuravlev wrote: could you also send me your mballoc.c please? thanks, Alex Aaron Knister wrote: Thanks so much for looking into this. Here's what I got from dmesg. Interestingly enough every time it panics the CPU listed is #4...do you think that points to a hardware problem? - --- [cut here ] - [please bite here ] - Kernel BUG at ...build/BUILD/lustre-ldiskfs-3.0.4/ldiskfs/mballoc.c:1334 invalid opcode: [1] SMP last sysfs file: /devices/pci:00/:00:02.0/:01:00.0/:02:02.0/:04:00.1/irq CPU 4 Modules linked in: obdfilter(U) fsfilt_ldiskfs(U) ost(U) mgc(U) ldiskfs(U) crc16(U) lustre(U) lov(U) lquota(U) mdc(U) ko2iblnd(U) ptlrpc(U) obdclass( U) lnet(U) lvfs(U) libcfs(U) netconsole(U) autofs4(U) hidp(U) rfcomm(U) l2cap(U) bluetooth(U) sunrpc(U) ip6t_REJECT(U) xt_tcpudp(U) ip6table_filter(U ) ip6_tables(U) x_tables(U) ipv6(U) ib_iser(U) libiscsi(U) scsi_transport_iscsi(U) rdma_ucm(U) ib_ucm(U) ib_srp(U) ib_sdp(U) rdma_cm(U) ib_cm(U) iw_c m(U) ib_addr(U) ib_local_sa(U) ib_ipoib(U) ib_sa(U) ib_uverbs(U) ib_umad(U) dm_multipath(U) video(U) sbs(U) backlight(U) i2c_ec(U) button(U) battery( U) asus_acpi(U) acpi_memhotplug(U) ac(U) parport_pc(U) lp(U) parport(U) ib_mthca(U) ib_mad(U) i2c_i801(U) e1000(U) ide_cd(U) ib_core(U) shpchp(U) i2c _core(U) cdrom(U) sg(U) pcspkr(U) dm_snapshot(U) dm_zero(U) dm_mirror(U) dm_mod(U) ata_piix(U) ahci(U) megaraid_sas(U) sata_sil(U) libata(U) sd_mod(U ) scsi_mod(U) ext3(U) jbd(U) ehci_hcd(U) ohci_hcd(U) uhci_hcd(U) Pid: 5717, comm: ll_ost_io_11 Tainted: GF 2.6.18-53.1.13.el5_lustre.1.6.4.3smp #1 RIP: 0010:[8885a01f] [8885a01f] :ldiskfs:ldiskfs_mb_use_best_found+0xef/0x520 RSP: 0018:81040e4dd320 EFLAGS: 00010246 RAX: RBX: 81040e4dd3d0 RCX: 007f RDX: 810406833000 RSI: 8103a825 RDI: 1000 RBP: 0800 R08: 0020 R09: 0010 R10: 81037d136ff8 R11: 8000 R12: 81040e4dd470 R13: 0001 R14: 0020 R15: 0800 FS: 2aac0220() GS:81042fc20b40() knlGS: CS: 0010 DS: ES: CR0: 8005003b CR2: 2e2ad0a0 CR3: 00041c6d1000 CR4: 06e0 Process ll_ost_io_11 (pid: 5717, threadinfo 81040e4dc000, task 81041228a040) Stack: 81040e4dd4c0 81040e4dd470 000b 0001 0001 0020 81040e4dd4c0 8885cd4e 81038c2b0a38 0010 81040e4dd4a0 Call Trace: [8885cd4e] :ldiskfs:ldiskfs_mb_regular_allocator+0x66e/0xd50 [888443c2] :ldiskfs:ldiskfs_mark_inode_dirty+0x132/0x160 [88859664] :ldiskfs:ldiskfs_mb_initialize_context+0x144/0x160 [8885e1e6] :ldiskfs:ldiskfs_mb_new_blocks+0x166/0x280 [888bf091] :fsfilt_ldiskfs:ldiskfs_ext_new_extent_cb+0x301/0x640 [8000a63f] get_page_from_freelist+0x223/0x3cf [882390b0] :ib_mthca:mthca_arbel_post_send+0x5a2/0x5b4 [88856d27] :ldiskfs:ldiskfs_ext_walk_space+0x1b7/0x250 [888bed90] :fsfilt_ldiskfs:ldiskfs_ext_new_extent_cb+0x0/0x640 [80054ccc] mwait_idle+0x0/0x4a [888ba9ae] :fsfilt_ldiskfs:fsfilt_map_nblocks+0xfe/0x150 [888eee08] :obdfilter:filter_direct_io+0x478/0xce0 [888f14de] :obdfilter:filter_commitrw_write+0x184e/0x2570 [8005a2ea] cache_alloc_refill+0x106/0x186 [8003ce86] lock_timer_base+0x1b/0x3c [8889eb46] :ost:ost_brw_write+0x21b6/0x28c0 [80088431] default_wake_function+0x0/0xe [888a283e] :ost:ost_handle+0x2a8e/0x58d8 [88603c82] :obdclass:class_handle2object+0xd2/0x160 [8869f230] :ptlrpc:lustre_swab_ptlrpc_body+0x0/0x90 [8869cde5] :ptlrpc:lustre_swab_buf+0xc5/0xf0 [886a4a3b] :ptlrpc:ptlrpc_server_handle_request+0xb0b/0x1270 [80060f29] thread_return+0x0/0xeb [8006b6c9] do_gettimeofday+0x50/0x92 [8855d056]
[Lustre-discuss] MDS crash and the Dilger Procedure
Hi, I just had to go through the Dilger procedure after MDS crashing when mounting MDT. The system is running fine now - glad that I just learned about this procedure. trace attached. [ 78.216468] Lustre: OBD class driver, [EMAIL PROTECTED] [ 78.217547] Lustre Version: 1.6.4.3 [ 78.218393] Build Version: 1.6.4.3-1970010101-PRISTINE-.home.goldbach.Build-lustre-server-kernel.linux-2.6.18.8-2.6.18.8-bnx2-1.6.7b-cciss-3.6.18-5-lustre-1.6.4.3 [ 78.299294] Lustre: Added LNI [EMAIL PROTECTED] [8/256] [ 78.300363] Lustre: Accept secure, port 988 [ 78.393841] Lustre: Lustre Client File System; [EMAIL PROTECTED] [ 78.453938] kjournald starting. Commit interval 5 seconds [ 78.461664] LDISKFS FS on dm-0, internal journal [ 78.462675] LDISKFS-fs: recovery complete. [ 78.469931] LDISKFS-fs: mounted filesystem with ordered data mode. [ 78.529838] kjournald starting. Commit interval 5 seconds [ 78.536134] LDISKFS FS on dm-0, internal journal [ 78.537334] LDISKFS-fs: mounted filesystem with ordered data mode. [ 78.594288] Lustre: MGS MGS started [ 78.595297] Lustre: Server MGS on device /dev/lustre_pool/mgt has started [ 78.676788] kjournald starting. Commit interval 5 seconds [ 78.681816] LDISKFS FS on dm-1, internal journal [ 78.682987] LDISKFS-fs: recovery complete. [ 78.688720] LDISKFS-fs: mounted filesystem with ordered data mode. [ 78.757422] kjournald starting. Commit interval 5 seconds [ 78.765206] LDISKFS FS on dm-1, internal journal [ 78.766255] LDISKFS-fs: mounted filesystem with ordered data mode. [ 78.864505] Lustre: Enabling user_xattr [ 78.907525] Lustre: 1361:0:(mds_fs.c:446:mds_init_server_data()) RECOVERY: service iloapp3-MDT, 3 recoverable clients, last_transno 158566404 [ 78.952173] Lustre: MDT iloapp3-MDT now serving dev (iloapp3-MDT/1455817d-55a2-a694-4403-7abfdae1606f), but will be in recovery until 3 clients reconnect, or if no clients reconnect for 4:10; during that time new clients will not be allowed to connect. Recovery progress can be monitored by watching /proc/fs/lustre/mds/iloapp3-MDT/recovery_status. [ 78.978179] Lustre: 1361:0:(mds_lov.c:858:mds_notify()) MDS iloapp3-MDT: in recovery, not resetting orphans on iloapp3-OST_UUID [ 79.012936] BUG: scheduling while atomic: mount.lustre/0x8101/1361 [ 79.014247] [ 79.014247] Call Trace: [ 79.015240] [8025973a] __sched_text_start+0x7a/0x769 [ 79.016616] [8023ab95] lock_timer_base+0x1b/0x3c [ 79.017943] [8022f226] del_timer+0x4e/0x57 [ 79.018971] [802130a8] sync_buffer+0x0/0x3f [ 79.020003] [8025a59a] io_schedule+0x28/0x34 [ 79.021235] [802130e3] sync_buffer+0x3b/0x3f [ 79.022490] [8025a8f5] __wait_on_bit+0x40/0x6f [ 79.023780] [802130a8] sync_buffer+0x0/0x3f [ 79.024955] [8025a990] out_of_line_wait_on_bit+0x6c/0x78 [ 79.026373] [80286a91] wake_bit_function+0x0/0x23 [ 79.027743] [80222c9f] __bread+0x62/0x77 [ 79.028724] [88318de2] :ldiskfs:read_block_bitmap+0xa2/0xf0 [ 79.030245] [88319695] :ldiskfs:ldiskfs_free_blocks_sb+0x115/0x510 [ 79.031856] [88319b21] :ldiskfs:ldiskfs_free_blocks+0x91/0xe0 [ 79.033392] [8831ed1a] :ldiskfs:ldiskfs_free_data+0x8a/0x110 [ 79.034944] [8831f19c] :ldiskfs:ldiskfs_truncate+0x20c/0x650 [ 79.036261] [802dbeab] start_this_handle+0x355/0x405 [ 79.037705] [8831fbb4] :ldiskfs:ldiskfs_delete_inode+0x84/0xf0 [ 79.039043] [8831fb30] :ldiskfs:ldiskfs_delete_inode+0x0/0xf0 [ 79.040636] [8022c804] generic_delete_inode+0x8e/0x10b [ 79.042042] [883ce891] :mds:mds_obd_destroy+0xa11/0xad0 [ 79.043435] [8022a2d7] mntput_no_expire+0x19/0x8b [ 79.044727] [880f361b] :obdclass:llog_lvfs_close+0x6b/0x130 [ 79.046201] [880f46c1] :obdclass:llog_lvfs_destroy+0x841/0xa10 [ 79.047756] [880f0a0f] :obdclass:llog_cat_id2handle+0x4cf/0x5f0 [ 79.049289] [8021557d] cache_grow+0x2ee/0x343 [ 79.050501] [880fa9c5] :obdclass:cat_cancel_cb+0x405/0x630 [ 79.051984] [880f0129] :obdclass:llog_process+0xa09/0xe20 [ 79.053398] [8020c894] dput+0x23/0x152 [ 79.054539] [880fa5c0] :obdclass:cat_cancel_cb+0x0/0x630 [ 79.055959] [880fa3b3] :obdclass:llog_obd_origin_setup+0x773/0x980 [ 79.057507] [8021819f] vsnprintf+0x55e/0x5a3 [ 79.058716] [880fb37d] :obdclass:llog_setup+0x78d/0x860 [ 79.060114] [8840ea94] :osc:osc_llog_init+0x104/0x390 [ 79.061491] [880f9855] :obdclass:__llog_ctxt_put+0x25/0xe0 [ 79.062957] [880f8979] :obdclass:obd_llog_init+0x179/0x210 [ 79.064362] [882632ca] :lov:lov_llog_init+0x2ca/0x400 [ 79.065556] [880f8979] :obdclass:obd_llog_init+0x179/0x210 [ 79.067021] [8022a2d7] mntput_no_expire+0x19/0x8b [
[Lustre-discuss] lustre and multi path
Our new lustre hardware arrived from sun today. Looking at the duel MDS and FC disk array for it. We will need multipath. Has anyone ever used multipath with lustre? Is there any issues? If we set up regular multipath via LVM lustre won't care as far as I can tell and browsing archives. What about multipath without LVM? Our StorageTek array has dual controllers with dual ports going to dual port FC cards in the MDS's. Each MDS has a connection to both controllers so we will need multipath to get any advantage to this. Comments? Brock Palen www.umich.edu/~brockp Center for Advanced Computing [EMAIL PROTECTED] (734)936-1985 ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] lustre and multi path
Hi Brock, We have sun oss's mds's and fc attached arrays all connected via an fc switch. We use sun's rdac driver rdac-LINUX-09.01.B2.74-source.tar. I think we tried to get the native RHEL4 multipath working but did not succeed with our configuration. Stuart On Thu, Jun 5, 2008 at 3:57 PM, Brock Palen [EMAIL PROTECTED] wrote: Our new lustre hardware arrived from sun today. Looking at the duel MDS and FC disk array for it. We will need multipath. Has anyone ever used multipath with lustre? Is there any issues? If we set up regular multipath via LVM lustre won't care as far as I can tell and browsing archives. What about multipath without LVM? Our StorageTek array has dual controllers with dual ports going to dual port FC cards in the MDS's. Each MDS has a connection to both controllers so we will need multipath to get any advantage to this. Comments? Brock Palen www.umich.edu/~brockp http://www.umich.edu/%7Ebrockp Center for Advanced Computing [EMAIL PROTECTED] (734)936-1985 ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] MDS crash and the Dilger Procedure
On Jun 05, 2008 23:28 +0200, Jakob Goldbach wrote: I just had to go through the Dilger procedure after MDS crashing when mounting MDT. The system is running fine now - glad that I just learned about this procedure. Hmm, looking at the size of this stack, I wonder if it is a stack overflow problem related to bug 15575? That was fixed in 1.6.5. At least the second stack looks to be related, and the first one is very deep... It does also appear that it is doing a reschedule while atomic in the first stack, though given the number of functions there it will be hard to determine what is making this atomic... Boy, do I love Linux stack traces with every function ever called on the stack... [ 79.012936] BUG: scheduling while atomic: mount.lustre/0x8101/1361 [ 79.014247] [ 79.014247] Call Trace: [ 79.015240] [8025973a] __sched_text_start+0x7a/0x769 [ 79.016616] [8023ab95] lock_timer_base+0x1b/0x3c [ 79.017943] [8022f226] del_timer+0x4e/0x57 [ 79.018971] [802130a8] sync_buffer+0x0/0x3f [ 79.020003] [8025a59a] io_schedule+0x28/0x34 [ 79.021235] [802130e3] sync_buffer+0x3b/0x3f [ 79.022490] [8025a8f5] __wait_on_bit+0x40/0x6f [ 79.023780] [802130a8] sync_buffer+0x0/0x3f [ 79.024955] [8025a990] out_of_line_wait_on_bit+0x6c/0x78 [ 79.026373] [80286a91] wake_bit_function+0x0/0x23 [ 79.027743] [80222c9f] __bread+0x62/0x77 [ 79.028724] [88318de2] :ldiskfs:read_block_bitmap+0xa2/0xf0 [ 79.030245] [88319695] :ldiskfs:ldiskfs_free_blocks_sb+0x115/0x510 [ 79.031856] [88319b21] :ldiskfs:ldiskfs_free_blocks+0x91/0xe0 [ 79.033392] [8831ed1a] :ldiskfs:ldiskfs_free_data+0x8a/0x110 [ 79.034944] [8831f19c] :ldiskfs:ldiskfs_truncate+0x20c/0x650 [ 79.036261] [802dbeab] start_this_handle+0x355/0x405 [ 79.037705] [8831fbb4] :ldiskfs:ldiskfs_delete_inode+0x84/0xf0 [ 79.039043] [8831fb30] :ldiskfs:ldiskfs_delete_inode+0x0/0xf0 [ 79.040636] [8022c804] generic_delete_inode+0x8e/0x10b [ 79.042042] [883ce891] :mds:mds_obd_destroy+0xa11/0xad0 [ 79.044727] [880f361b] :obdclass:llog_lvfs_close+0x6b/0x130 The fput-dput-iput-generic_drop_inode() part of the chain _appears_ to be where the stack is going, and following it down further ldiskfs_delete_inode- ldiskfs_truncate- ldiskfs_free_data- ldiskfs_free_blocks- ldiskfs_free_blocks_sb- read_block_bitmap- __bread() is calling might_sleep(), so the fact that we schedule later on shouldn't come as a surprise, unless, it seems, the kernel is compiled without CONFIG_DEBUG_SPINLOCK_SLEEP enabled. In that case, this kind of problem would be hit only when we actually DO sleep waiting for IO, [ 79.046201] [880f46c1] :obdclass:llog_lvfs_destroy+0x841/0xa10 It seems we are going into llog_lvfs_destroy() while processing the configuration log, which is a bit strange, unless this is due to the MDS-LOV-OSC hitting an empty log during startup and deleting it. That in itself shouldn't be harmful... There is (as can be seen below) a twisty maze of callbacks related to the distributed logging code, but it doesn't appear to be holding a spinlock that would make it atomic. Looking at my kernel, in_atomic() is set for a number of reasons (IRQ, NMI, etc) but the most common reason is because of spinlock, unless CONFIG_PREEMPT is enabled. [ 79.047756] [880f0a0f] :obdclass:llog_cat_id2handle+0x4cf/0x5f0 [ 79.050501] [880fa9c5] :obdclass:cat_cancel_cb+0x405/0x630 [ 79.051984] [880f0129] :obdclass:llog_process+0xa09/0xe20 In 1.6.5 (due to bug 15575) a new thread is started at this point to process the per-OSC log. [ 79.054539] [880fa5c0] :obdclass:cat_cancel_cb+0x0/0x630 [ 79.055959] [880fa3b3] :obdclass:llog_obd_origin_setup+0x773/0x980 [ 79.058716] [880fb37d] :obdclass:llog_setup+0x78d/0x860 [ 79.060114] [8840ea94] :osc:osc_llog_init+0x104/0x390 [ 79.061491] [880f9855] :obdclass:__llog_ctxt_put+0x25/0xe0 [ 79.062957] [880f8979] :obdclass:obd_llog_init+0x179/0x210 [ 79.064362] [882632ca] :lov:lov_llog_init+0x2ca/0x400 [ 79.065556] [880f8979] :obdclass:obd_llog_init+0x179/0x210 [ 79.068313] [883941ad] :mds:mds_llog_init+0x1ad/0x270 [ 79.070777] [880f8979] :obdclass:obd_llog_init+0x179/0x210 [ 79.073520] [880f8dc5] :obdclass:llog_cat_initialize+0x3b5/0x670 [ 79.075090] [88277c61] :lov:lov_get_info+0x9f1/0xaa0 [ 79.076405] [8828f6c1] :lov:qos_add_tgt+0x681/0x730 [ 79.077573] [8839d5ac] :mds:mds_lov_update_desc+0xbcc/0xd30 [ 79.079050] [883a03ec] :mds:mds_notify+0x36c/0x690 [ 79.080630] [8826c657] :lov:lov_notify+0xa97/0xfd0 [ 79.081820]
Re: [Lustre-discuss] lustre and multi path
Hi Brock, I've got a Sun StorageTek array hooked up to one of our clusters, and I'm using labels instead of multi-pathing. We've got it hooked up in a similar fashion as Stuart; it's a bit slow and sloppy when initializing, but it works well enough and there are no problems once OSTs are online. Klaus On 6/5/08 3:57 PM, Brock Palen [EMAIL PROTECTED]did etch on stone tablets: Our new lustre hardware arrived from sun today. Looking at the duel MDS and FC disk array for it. We will need multipath. Has anyone ever used multipath with lustre? Is there any issues? If we set up regular multipath via LVM lustre won't care as far as I can tell and browsing archives. What about multipath without LVM? Our StorageTek array has dual controllers with dual ports going to dual port FC cards in the MDS's. Each MDS has a connection to both controllers so we will need multipath to get any advantage to this. Comments? Brock Palen www.umich.edu/~brockp Center for Advanced Computing [EMAIL PROTECTED] (734)936-1985 ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] lustre and multi path
Hi Brock, Yeah, that's likely to be an issue if each host has more than one path ... What about using HA to force one path to be inactive at the device level? I know QLogic FC cards support this functionality, although it requires changing the options used by the driver kernel module ... mind you, comparing that to a solution using the multipath daemon, that's six of one, half a dozen of the other, I'd think. Klaus On 6/5/08 5:09 PM, Brock Palen [EMAIL PROTECTED]did etch on stone tablets: This would be for the MDS/MGS only, but thats good to know. Problem is our two MDS servers (active/passive) will have two connections each to the same lun, so there could be issues. Brock Palen www.umich.edu/~brockp Center for Advanced Computing [EMAIL PROTECTED] (734)936-1985 On Jun 5, 2008, at 7:52 PM, Klaus Steden wrote: Hi Brock, I've got a Sun StorageTek array hooked up to one of our clusters, and I'm using labels instead of multi-pathing. We've got it hooked up in a similar fashion as Stuart; it's a bit slow and sloppy when initializing, but it works well enough and there are no problems once OSTs are online. Klaus On 6/5/08 3:57 PM, Brock Palen [EMAIL PROTECTED]did etch on stone tablets: Our new lustre hardware arrived from sun today. Looking at the duel MDS and FC disk array for it. We will need multipath. Has anyone ever used multipath with lustre? Is there any issues? If we set up regular multipath via LVM lustre won't care as far as I can tell and browsing archives. What about multipath without LVM? Our StorageTek array has dual controllers with dual ports going to dual port FC cards in the MDS's. Each MDS has a connection to both controllers so we will need multipath to get any advantage to this. Comments? Brock Palen www.umich.edu/~brockp Center for Advanced Computing [EMAIL PROTECTED] (734)936-1985 ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss