February 14, 201214 yr Is this indicating that the kernel is running out of memory? I have 8GB so hope not, and only run unMenu, SABnzbd. Feb 11 23:52:24 RCNAS kernel: swapper: page allocation failure. order:0, mode:0x4020 Feb 11 23:52:24 RCNAS kernel: Pid: 0, comm: swapper Not tainted 2.6.32.9-unRAID #8 (Errors) Feb 11 23:52:24 RCNAS kernel: Call Trace: (Errors) Feb 11 23:52:24 RCNAS kernel: [<c104d06f>] __alloc_pages_nodemask+0x3fb/0x42f (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1258980>] ? tcp_v4_do_rcv+0x21/0x162 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c106841c>] __slab_alloc+0x13e/0x425 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1068d8f>] __kmalloc_track_caller+0x86/0xcf (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1229a38>] ? __netdev_alloc_skb+0x17/0x34 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1229a38>] ? __netdev_alloc_skb+0x17/0x34 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1229709>] __alloc_skb+0x50/0x119 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1229a38>] __netdev_alloc_skb+0x17/0x34 (Errors) Feb 11 23:52:24 RCNAS kernel: [<f8632dfb>] e1000_alloc_rx_buffers+0x79/0x1bf [e1000e] (Errors) Feb 11 23:52:24 RCNAS kernel: [<f8632a24>] e1000_clean_rx_irq+0x259/0x2cb [e1000e] (Errors) Feb 11 23:52:24 RCNAS kernel: [<f86318de>] e1000_clean+0x62/0x1de [e1000e] (Errors) Feb 11 23:52:24 RCNAS kernel: [<c12312aa>] net_rx_action+0x57/0x102 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1028261>] __do_softirq+0x84/0xf8 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c10282fb>] do_softirq+0x26/0x2b (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1028556>] irq_exit+0x29/0x2b (Errors) Feb 11 23:52:24 RCNAS kernel: [<c10042c5>] do_IRQ+0x80/0x96 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1002f29>] common_interrupt+0x29/0x30 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c116d019>] ? acpi_idle_enter_bm+0x216/0x23f (Errors) Feb 11 23:52:24 RCNAS kernel: [<c12108e8>] cpuidle_idle_call+0x63/0x9b (Errors) Feb 11 23:52:24 RCNAS kernel: [<c1001a14>] cpu_idle+0x3a/0x4e (Errors) Feb 11 23:52:24 RCNAS kernel: [<c128a8bf>] rest_init+0x53/0x55 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c13f580c>] start_kernel+0x27b/0x280 (Errors) Feb 11 23:52:24 RCNAS kernel: [<c13f5091>] i386_start_kernel+0x91/0x96 (Errors) Full syslog attached. (missing lines are just the move logger entries). Any ideas what is causing this? Do I need to worry? Everything seems to be running fine otherwise. syslog-2012-02-14.zip
February 15, 201214 yr Are you running cache_dirs ? How many files do you have across all your drives? Is /tmp or /var/tmp or any of the /var/spool directories filling up? How large does syslog get?
February 16, 201214 yr Author no 'cache_dirs' - thought about it, but decided against it just to be safe as I'm running SABnzbd and it could take a bit of memory when doing verifications/unrar'ing. which I don't believe was happening at the time of those messages. only 288, 16 and 8 bytes in each of those folders you listed.
February 16, 201214 yr no 'cache_dirs' - thought about it, but decided against it just to be safe as I'm running SABnzbd and it could take a bit of memory when doing verifications/unrar'ing. which I don't believe was happening at the time of those messages. only 288, 16 and 8 bytes in each of those folders you listed. what do you get when you type: free
February 16, 201214 yr Author total used free shared buffers cached Mem: 8300336 8028104 272232 0 98148 7429296 -/+ buffers/cache: 500660 7799676 Swap: 0 0 0
February 16, 201214 yr The issue is low memory. Otherwise you would not run out. You could try adding swap space. Or just a swap file on loopback. This has always been my beef with initramfs. It cannot be swapped out. If it were on tmpfs then it could be swapped. Come to think of it, Where is your applications tmp folder set to? you may want to mount a tmpfs filesystem on /tmp or /var/tmp or redirect them with a mount -o bind to a folder on your cache drive mount -t tmpfs -o mode=1777 tmpfs /tmp This will set it to 50% of available ram. You can specify the size. If you have a swap file and the system needs ram, while tmp is full, it will swap out unused portions of /tmp. So if you had an 8gb swap partition or file and set /tmp with tmpfs of 8G it would swap out if the kernel became pressured for ram. Also, Do an lsof and inspect what files might still be open.
February 16, 201214 yr You should never run out of memory with just unMENU and SABnzbd running. I run that plus other stuff on 2gig of memory with no problems. Make sure you're not using the RAM filesystem for some type of file storage. Peter
Archived
This topic is now archived and is closed to further replies.