[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| Title: | DIGITAL UNIX (FORMERLY KNOWN AS DEC OSF/1) |
| Notice: | Welcome to the Digital UNIX Conference |
| Moderator: | SMURF::DENHAM |
|
| Created: | Thu Mar 16 1995 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 10068 |
| Total number of notes: | 35879 |
Can anybody explain following stack-trace:
(dbx) set $pid=10681
(dbx) t
> 0 thread_block() ["../../../../src/kernel/kern/sched_prim.c":1932,
0xfffffc00004766e4]
1 thread_sleep(event = 18446739675667068816, lock = 0xfffffc0000699f60,
interruptible = 4013156)]
2 _cond_wait(0xffffffff806058f0, 0xbb52, 0xff6, 0x0, 0xfffffc00003d381c)
["../../../../src/kerne]
3 _state_block(bp = 0xffffffffa2312fc8, wait = 0x1)
["../../../../src/kernel/msfs/bs/bs_buffer2.]
4 bs_refpg_int(bfPageRefH = 0xffffffffa3b9b480, bfPageAddr =
0xffffffffa3b9b478, bfAccessH = 274]
5 bs_refpg(bfPageRefH = 0xffffffffa3b9b480, bfPageAddr = 0x5df,
bfAccessH = 446, bsPage = 48982,]
6 fs_read(0xfffffc00286f1000, 0xffffffffa3b9b768, 0x17eac000, 0x1000,
0xfffffc0000000000) ["../.]
7 msfs_read(vp = 0xfffffc00286f1000, uio = 0xffffffffa3b9b768, ioflag =
0, cred = 0xfffffc002b20]
8 vn_read(0xffffffffa3b9b708, 0xffffffffa3b9b768, 0xfffffc002b205000,
0x18ba9000, 0xfffffc000045]
More (n if no)?
9 rwuio(0xfffffc002bbc2210, 0x0, 0xffffffffa3b9b8b8, 0xffffffffa3b9b768,
0x0) ["../../../../src/]
10 read(0xfffffc0000001000, 0xffffffffa3b9b758, 0xc512000,
0x100000000001, 0x0) ["../../../../src]
11 syscall(0x12025c88c, 0x8, 0xfffffc003b500180, 0xc, 0x3)
["../../../../src/kernel/arch/alpha/sy]
12 _Xsyscall(0x8, 0x1207d4798, 0x14003aaf0, 0x8, 0x590d000)
["../../../../src/kernel/arch/alpha/l]
I have a customer with performance problems with an oracle-application.
The process from which i have taken this trace is one of two oracle-daemons
(the application accesses two databases), and is very often in this state
during the application-run. The 2nd oracle-background process doen't show
this behavior. May be, this _state_block is one of the reasons of bad
performance. During the run the cpu is 50 - 60 % idle, 200 MB memory free,
and i cannot detect physical i-o bottlenecks.
Digital-Unix V3.2d-1, Adfvs-Patches installed, 2 KZPSA-Adapters (2 HSZ40).
Turbolaser 8200, 2 CPU's.
| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 10001.1 | | DECWET::MARTIN | | Fri May 30 1997 15:39 | 8 |
| At a guess, your problem is that AdvFS is funneled (single-CPU only) in v3.2*.
You'll have to upgrade to v4.0 to get an SMP-enabled (multi-processor) AdvFS.
Note that this is only a guess, as I've never seen this particular stack trace
before. My only clues are that you're running 3.2d-1, you have 2 CPUs, and that
levels 3 - 7 in the stack trace are within AdvFS kernel code.
--Ken
|