Petr Vacek
2011-Jan-19 23:32 UTC
[Xen-users] xen-4.0.1 from make world - hypervisor alone consumes around 800 MB RAM , OOMs when try to baloon to lower amount, no domUs running yet
Greetings, I''ve just setuped a new AMD opteron 6168 12-core server with Xen-4.0.1 - downloading sources , running ''make world'' - adding raid+ata modules into kernel and running ... Everything works fine (filesystem,raid,domUs crunching numbers) , save for one troubling memory issue with dom0 : When running with hypervisor enabled , I see used memory as 860 MB (in top) right after linux start, even without Xend service and any domU running, when I strech domain 0 under 800 MB , I get OOM killer and restart. This way I know that I cannot possibly use this memory and it''s eaten up - I cannot give it to domU certainly . I can see this amount in meminfo in directMap4k , but it does not add up with other parts of meminfo to me - it does not belong to any visible process. When I use same booting configuration without Xen hypervisor, this memory usage goes away - linux eats about 60 MB , only service running is SSH and very basic Ubuntu 10.10. Kernel is from xen sources : 2.6.32.27 - kernel config as here ( http://pavouk.mediawork.cz/xen/2011/kernel-config.txt ) I''ve placed dmesg, xm dmesg, dmidecode machine information here: http://pavouk.mediawork.cz/xen/2011/ I''ve tried building 4.0.2 - memory issue was still present + kernel bugs so I reverted to older 4.0.1 runnning otherwise stable for my setup. Could this be caused by some kernel misconfiguration or xen hypervisor misconfiguration ? I''ve noticed some warning during initialization , but I think it''s something different : Performance Events: AMD PMU driver. [ 0.006998] ------------[ cut here ]------------ [ 0.006998] WARNING: at /opt/xen/xen-4.0.1/linux-2.6-pvops.git/arch/x86/xen/enlighten.c:729 xen_apic_write+0x15/0x17() [ 0.006998] Hardware name: H8SGL [ 0.006998] Modules linked in: [ 0.006998] Pid: 0, comm: swapper Not tainted 2.6.32.27 #9 [ 0.006998] Call Trace: [ 0.006998] [<ffffffff8107970b>] warn_slowpath_common+0x77/0x8f [ 0.006998] [<ffffffff81079732>] warn_slowpath_null+0xf/0x11 [ 0.006998] [<ffffffff810355c2>] xen_apic_write+0x15/0x17 [ 0.006998] [<ffffffff8104872a>] perf_events_lapic_init+0x2e/0x30 [ 0.006998] [<ffffffff81a976a2>] init_hw_perf_events+0x33e/0x3db [ 0.006998] [<ffffffff81a971c0>] identify_boot_cpu+0x3c/0x3e [ 0.006998] [<ffffffff81a97340>] check_bugs+0x9/0x2d [ 0.006998] [<ffffffff81a8ddbf>] start_kernel+0x432/0x447 [ 0.006998] [<ffffffff81a8d2c1>] x86_64_start_reservations+0xac/0xb0 [ 0.006998] [<ffffffff81a91ad0>] xen_start_kernel+0x5b7/0x5be [ 0.006998] ---[ end trace 4eaa2a86a8e2da22 ]--- The troubling meminfo dump: cat /proc/meminfo : MemTotal: 1229376 kB MemFree: 345440 kB Buffers: 412 kB Cached: 8572 kB SwapCached: 0 kB Active: 45304 kB Inactive: 14112 kB Active(anon): 38464 kB Inactive(anon): 12304 kB Active(file): 6840 kB Inactive(file): 1808 kB Unevictable: 20 kB Mlocked: 20 kB SwapTotal: 0 kB SwapFree: 0 kB Dirty: 948 kB Writeback: 0 kB AnonPages: 50508 kB Mapped: 6600 kB Shmem: 320 kB Slab: 15904 kB SReclaimable: 5236 kB SUnreclaim: 10668 kB KernelStack: 1016 kB PageTables: 1664 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 614688 kB Committed_AS: 153440 kB VmallocTotal: 34359738367 kB VmallocUsed: 264292 kB VmallocChunk: 34359473836 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 2048 kB DirectMap4k: 8647280 kB DirectMap2M: 0 kB Thanks in advance for any hints how to tamper this memory burden Regards Petr Vacek _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users