Re: [opensuse] Processes are slow on opensuse 11.3
Am Montag, 22. November 2010 schrieben Sie:
Am 22.11.2010 09:37, schrieb Per Jessen:
Christoph Bartoschek wrote:
Hi,
here is the output of top sorted by memory consumption. All the chipbench processes are cpu-bound. I wonder why 160 GB are cached while 56 GB from the processes are in swap.
Your chipbench processes seem to be mostly sleeping (S) or waiting for IO (D) - have you looked at what they are waiting for?
Yes. iotop shows that most of the time they are paging in. They want to read their memory. But it seems that the pages do not stay in RAM.
/proc/meminfo says that the memory is not stuck in page cache. It is somehow used by slab. What is this and how to get its usage down? MemTotal: 198493288 kB MemFree: 853372 kB Buffers: 824 kB Cached: 26108 kB SwapCached: 6369336 kB Active: 37073576 kB Inactive: 1104932 kB Active(anon): 37059712 kB Inactive(anon): 1090980 kB Active(file): 13864 kB Inactive(file): 13952 kB Unevictable: 0 kB Mlocked: 0 kB SwapTotal: 209713148 kB SwapFree: 149362056 kB Dirty: 16 kB Writeback: 0 kB AnonPages: 37642012 kB Mapped: 13312 kB Shmem: 0 kB Slab: 158765316 kB SReclaimable: 158732380 kB SUnreclaim: 32936 kB KernelStack: 2968 kB PageTables: 202500 kB NFS_Unstable: 4 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 308959792 kB Committed_AS: 64376360 kB VmallocTotal: 34359738367 kB VmallocUsed: 736572 kB VmallocChunk: 34358994676 kB Christoph -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Christoph Bartoschek wrote:
Am Montag, 22. November 2010 schrieben Sie:
Am 22.11.2010 09:37, schrieb Per Jessen:
Christoph Bartoschek wrote:
Hi,
here is the output of top sorted by memory consumption. All the chipbench processes are cpu-bound. I wonder why 160 GB are cached while 56 GB from the processes are in swap.
Your chipbench processes seem to be mostly sleeping (S) or waiting for IO (D) - have you looked at what they are waiting for?
Yes. iotop shows that most of the time they are paging in. They want to read their memory. But it seems that the pages do not stay in RAM.
/proc/meminfo says that the memory is not stuck in page cache. It is somehow used by slab. What is this and how to get its usage down?
I have seen a similar problem, but it's a couple of years ago - IIRC, mine was to do with with directories with large numbers of files (100.000s). What does /proc/slabinfo say? -- Per Jessen, Zürich (4.2°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Am Montag, 22. November 2010 schrieb Christoph Bartoschek:
Am Montag, 22. November 2010 schrieben Sie:
Am 22.11.2010 09:37, schrieb Per Jessen:
Christoph Bartoschek wrote:
Hi,
here is the output of top sorted by memory consumption. All the chipbench processes are cpu-bound. I wonder why 160 GB are cached while 56 GB from the processes are in swap.
Your chipbench processes seem to be mostly sleeping (S) or waiting for IO (D) - have you looked at what they are waiting for?
Yes. iotop shows that most of the time they are paging in. They want to read their memory. But it seems that the pages do not stay in RAM.
/proc/meminfo says that the memory is not stuck in page cache. It is somehow used by slab. What is this and how to get its usage down?
I found some information with slabtop: Active / Total Objects (% used) : 364597 / 1070670469 (0.0%) Active / Total Slabs (% used) : 52397 / 39688960 (0.1%) Active / Total Caches (% used) : 107 / 193 (55.4%) Active / Total Size (% used) : 159579.25K / 150697605.41K (0.1%) Minimum / Average / Maximum Object : 0.02K / 0.14K / 4096.00K OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME 1070187012 0 0% 0.14K 39636556 27 158546224K ext4_alloc_context Nearly the whole memory is held in ext4_alloc_context and not released properly for other processes. The question is now how to free this memory? Christoph -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Am Montag, 22. November 2010 schrieb Christoph Bartoschek:
Am Montag, 22. November 2010 schrieb Christoph Bartoschek:
Am Montag, 22. November 2010 schrieben Sie:
Am 22.11.2010 09:37, schrieb Per Jessen:
Christoph Bartoschek wrote:
Hi,
here is the output of top sorted by memory consumption. All the chipbench processes are cpu-bound. I wonder why 160 GB are cached while 56 GB from the processes are in swap.
Your chipbench processes seem to be mostly sleeping (S) or waiting for IO (D) - have you looked at what they are waiting for?
Yes. iotop shows that most of the time they are paging in. They want to read their memory. But it seems that the pages do not stay in RAM.
/proc/meminfo says that the memory is not stuck in page cache. It is somehow used by slab. What is this and how to get its usage down?
I found some information with slabtop:
Active / Total Objects (% used) : 364597 / 1070670469 (0.0%) Active / Total Slabs (% used) : 52397 / 39688960 (0.1%) Active / Total Caches (% used) : 107 / 193 (55.4%) Active / Total Size (% used) : 159579.25K / 150697605.41K (0.1%) Minimum / Average / Maximum Object : 0.02K / 0.14K / 4096.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME 1070187012 0 0% 0.14K 39636556 27 158546224K ext4_alloc_context
Nearly the whole memory is held in ext4_alloc_context and not released properly for other processes.
The strange thing is that all of our data is held in NFS. ext4 is only used for the base installation. Christoph -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Christoph Bartoschek wrote:
Am Montag, 22. November 2010 schrieb Christoph Bartoschek:
I found some information with slabtop:
Active / Total Objects (% used) : 364597 / 1070670469 (0.0%) Active / Total Slabs (% used) : 52397 / 39688960 (0.1%) Active / Total Caches (% used) : 107 / 193 (55.4%) Active / Total Size (% used) : 159579.25K / 150697605.41K (0.1%) Minimum / Average / Maximum Object : 0.02K / 0.14K / 4096.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME 1070187012 0 0% 0.14K 39636556 27 158546224K ext4_alloc_context
Nearly the whole memory is held in ext4_alloc_context and not released properly for other processes.
The strange thing is that all of our data is held in NFS. ext4 is only used for the base installation.
Did you look at /proc/slabinfo ? -- Per Jessen, Zürich (4.1°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Am Montag, 22. November 2010 schrieb Per Jessen:
Christoph Bartoschek wrote:
Am Montag, 22. November 2010 schrieb Christoph Bartoschek:
I found some information with slabtop: Active / Total Objects (% used) : 364597 / 1070670469 (0.0%) Active / Total Slabs (% used) : 52397 / 39688960 (0.1%) Active / Total Caches (% used) : 107 / 193 (55.4%) Active / Total Size (% used) : 159579.25K / 150697605.41K (0.1%) Minimum / Average / Maximum Object : 0.02K / 0.14K / 4096.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
1070187012 0 0% 0.14K 39636556 27 158546224K ext4_alloc_context
Nearly the whole memory is held in ext4_alloc_context and not released properly for other processes.
The strange thing is that all of our data is held in NFS. ext4 is only used for the base installation.
Did you look at /proc/slabinfo ?
It shows the same information as slabtop. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (2)
-
Christoph Bartoschek
-
Per Jessen