using 2.3.42 for some hours I realized that the system was very low
on free memory or buffer space without any obvious reason (ps output below).
where are those 82MB `used' ?? it's not only that the output of `free'
might be bogus, I realized the problem because even small amount of files
haven't been cached any more when running the same `diff ....' command
several times:
-------------------------------------------------------------------------------
total used free shared buffers cached
Mem: 126596 111448 15148 0 1160 27980
-/+ buffers/cache: 82308 44288
Swap: 82364 0 82364
SysRq: Show Memory
Mem-info:
Free pages: 14756kB ( 0kB HighMem)
( Free: 3689, lru_cache: 7654 (256 512 768) )
DMA: 136*4kB 8*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB 1*512kB 1*1024kB 0*2048kB = 2640kB)
Normal: 1019*4kB 355*8kB 139*16kB 57*32kB 16*64kB 1*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB = 12116kB)
HighMem: = 0kB)
Swap cache: add 2276, delete 1787, find 674/3186
Free swap: 80128kB
32768 pages of RAM
0 pages of HIGHMEM
1119 reserved pages
8135 pages shared
489 pages swap cached
0 pages in page table cache
Buffer memory: 1096kB
Networking buffers in use : 32
Total network buffer allocations : 1873464
Total failed network buffer allocs : 0
IP fragment buffer size : 0
-------------------------------------------------------------------------------
only after running `eatmem 120M', a small tool to allocate/trash memory,
the situation cleared up and magically I've got a lot of free space again
and files got buffered again:
-------------------------------------------------------------------------------
total used free shared buffers cached
Mem: 126596 10652 115944 0 312 2344
-/+ buffers/cache: 7996 118600
Swap: 82364 0 82364
Mem-info:
Free pages: 90240kB ( 0kB HighMem)
( Free: 22560, lru_cache: 3383 (256 512 768) )
DMA: 35*4kB 30*8kB 32*16kB 32*32kB 23*64kB 20*128kB 11*256kB 6*512kB 2*1024kB 0*2048kB = 13884kB)
Normal: 5*4kB 2*8kB 2*16kB 34*32kB 281*64kB 163*128kB 84*256kB 21*512kB 2*1024kB 1*2048kB = 76356kB)
HighMem: = 0kB)
Swap cache: add 5082, delete 5082, find 1454/5998
Free swap: 82364kB
32768 pages of RAM
0 pages of HIGHMEM
1119 reserved pages
9360 pages shared
0 pages swap cached
0 pages in page table cache
Buffer memory: 880kB
Networking buffers in use : 35
Total network buffer allocations : 1885942
Total failed network buffer allocs : 0
IP fragment buffer size : 0
-------------------------------------------------------------------------------
I've used 2.3.38/39 for quite a while (but never .40 or .41)
and I've never seen such behaviour before.
here is the output of `ps' before using `eatmem':
-------------------------------------------------------------------------------
# ps axvOT
PID TTY STAT TIME MAJFL TRS DRS RSS %MEM COMMAND
1 ? S 0:15 139 330 41 204 0.1 init
2 ? SW 0:06 0 0 0 0 0.0 [kswapd]
3 ? SW 0:00 0 0 0 0 0.0 [kflushd]
4 ? SW 0:00 0 0 0 0 0.0 [kupdate]
5 ? SW 0:00 0 0 0 0 0.0 [md_thread]
45 ? SW 0:00 0 0 0 0 0.0 [scsi_eh_1]
100 ? S 0:00 33 23 1136 316 0.2 /sbin/portmap
104 ? S 0:00 14 23 1392 324 0.2 /usr/sbin/rpc.ugidd
112 ? S 0:00 99 27 1448 592 0.4 /usr/sbin/syslogd
116 ? S 0:00 27 19 1636 896 0.7 /usr/sbin/klogd -c 1
160 ? S 0:00 85 15 1492 588 0.4 automount -t 300 /misc f
169 ? S 0:00 32 60 1771 600 0.4 /usr/sbin/rpc.mountd
172 ? S 0:00 30 62 1785 600 0.4 /usr/sbin/rpc.nfsd
179 ? S 0:00 9 12 1255 332 0.2 /usr/sbin/atd
184 ttyS1 S 0:00 19 56 1115 348 0.2 /usr/sbin/gpm -t logi -b
188 ? S 0:00 66 15 1428 512 0.4 /usr/sbin/inetd
201 ? S 0:00 25 46 1133 340 0.2 /usr/sbin/lpd
215 ? S 0:00 129 46 1469 608 0.4 /usr/sbin/lpd
220 ? S 0:00 25 7 1424 412 0.3 /usr/sbin/rwhod
225 ? S 0:00 58 364 1875 864 0.6 sendmail: accepting conn
230 ? S 0:00 18 215 1492 536 0.4 /usr/sbin/sshd -n
234 ? SL 0:00 322 215 1440 1648 1.3 /usr/sbin/xntpd
238 ? S 0:00 55 18 1273 584 0.4 /usr/sbin/cron
242 ? S 0:00 12 24 1435 584 0.4 /usr/sbin/in.identd -e
243 ? S 0:00 1 24 1435 584 0.4 /usr/sbin/in.identd -e
244 ? S 0:00 0 24 1435 584 0.4 /usr/sbin/in.identd -e
245 ? S 0:00 0 24 1435 584 0.4 /usr/sbin/in.identd -e
246 ? S 0:00 0 24 1435 584 0.4 /usr/sbin/in.identd -e
247 ? S 0:00 0 24 1435 584 0.4 /usr/sbin/in.identd -e
248 ? S 0:00 0 24 1435 584 0.4 /usr/sbin/in.identd -e
257 ? S 0:00 43 30 1409 616 0.4 /usr/sbin/nscd
259 ? S 0:00 1 30 1409 616 0.4 /usr/sbin/nscd
260 ? S 0:00 2 30 1409 616 0.4 /usr/sbin/nscd
261 ? S 0:00 4 30 1409 616 0.4 /usr/sbin/nscd
262 ? S 0:00 0 30 1409 616 0.4 /usr/sbin/nscd
263 ? S 0:00 1 30 1409 616 0.4 /usr/sbin/nscd
264 ? S 0:00 2 30 1409 616 0.4 /usr/sbin/nscd
281 tty1 S 0:00 345 21 1694 1008 0.7 login -- root
282 tty2 S 0:00 104 8 1107 428 0.3 /sbin/mingetty tty2
283 tty3 S 0:00 104 8 1107 428 0.3 /sbin/mingetty tty3
284 tty4 S 0:00 104 8 1107 428 0.3 /sbin/mingetty tty4
285 tty5 S 0:00 104 8 1107 428 0.3 /sbin/mingetty tty5
286 tty6 S 0:00 104 8 1107 428 0.3 /sbin/mingetty tty6
287 ttyS4 S 0:00 174 29 1454 632 0.4 /usr/sbin/getty ttyS4 19
291 tty1 S 0:00 294 448 2003 1520 1.2 -bash
13018 tty1 T 0:00 154 76 1419 648 0.5 less
13024 tty1 R 0:00 230 77 2598 1004 0.7 ps axvOT
-------------------------------------------------------------------------------
Harald
--
All SCSI disks will from now on ___ _____
be required to send an email notice 0--,| /OOOOOOO\
24 hours prior to complete hardware failure! <_/ / /OOOOOOOOOOO\
\ \/OOOOOOOOOOOOOOO\
\ OOOOOOOOOOOOOOOOO|//
Harald Koenig, \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik // / \\ \
koenig@tat.physik.uni-tuebingen.de ^^^^^ ^^^^^
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Feb 07 2000 - 21:00:09 EST