From: REEVE Ruth (RREEVE@goldcoast.qld.gov.au)
Date: Fri May 31 2002 - 02:48:14 EDT
Thanks for all who replied (Phil Baldwin, Robert Mulley, Mathew Butler, Jim Lola & Kat),
and thanks to Mat for the neat script :) However it was mainly a case of playing with
the parameters according to oracle and o/s tuning manuals...........
It appears to have been a few of the tunable parameters, not sure
if they all combined to help, but this is what I did :-
I changed the following parameter as below (as indicated by a few people
who replied). However this on it's own didn't fix the problem.
ipc:
shm_max = 2139095040 (was 1073741824)
The DBA's found some more oracle info so I tried changing the following & rebooting
proc:
max_per_proc_address_space = 10737418240 (was 4294967296)
per_proc_address_space = 10737418240 (was 4294967296)
ipc:
shm_mni = 256 (was 1024)
shm_seg = 1024 (was 256)
sem_msl = 2600 (was 1024)
vm:
ubc_maxpercent = 70 (was 30)
but still got the same error (see below).....doh!!!!!!
Then I noticed something with the stack_sizes.......As the oracle doco said
not to set the stack_sizes larger than 500MB, and the doco we had
on system tuning said to "choose values that are significantly less that the
address space limit", I tried setting them to the default values to see what happened....
proc:
per_proc_stack_size = 8388608 (was 4398046511104)
max_per_proc_stack_size = 33554432 (was 4398046511104)
and guess what................it worked!!!!!!
yahoooooooooooooooooooooooooooooooooooooooooooooooooo
finally.....phew........
Note:
also....as I was going through the tuning doco for large memory systems, I set some parameters as follows to the recommended values to keep in sync
with the values set above, however this was done after the oracle error
was fixed above.
per_proc_data_size = 10737418240 (was 4398046511104)
max_per_proc_data_size = 10737418240 (was 4398046511104)
thankyou again
Ruth
> -----Original Message-----
> From: REEVE Ruth
> Sent: Friday, 31 May 2002 10:56 AM
> To: 'tru64-unix-managers@ornl.gov'
> Subject: Oracle 7.3.4 problem on Tru64 5.1A PK1
>
> Hi All,
>
> We have a new ES45 with 2 x EV6.8CB cpu's, 4GB memory, 12GB swap
> running Tru64 5.1A PK1 and using Advfs. The DBA's are trying to
> start Oracle 7.3.4 but get the following error :-
>
> uxfmis1 (root) #> su - oracle
>
> SIDs on this machine are
> mprd FPRD
>
> ORACLE_SID = [FPRD] ?
> FPRD@uxfmis1:> svrmgrl
>
> Oracle Server Manager Release 2.3.4.0.0 - Production
>
> Copyright (c) Oracle Corporation 1994, 1995. All rights reserved.
>
> Oracle7 Server Release 7.3.4.4.0 with the 64-bit option - Production
> With the distributed and parallel query options
> PL/SQL Release 2.3.4.4.0 - Production
>
> SVRMGR> connect internal
> Connected to an idle instance.
> SVRMGR> startup nomount
> ORA-07307: sms1sg: shmat error, unable to attach sga.
> Digital Unix Error: 12: Not enough space
> Additional information: 1
> SVRMGR> exit
> Server Manager complete.
>
>
> And yes, there is plenty of disk space :)
>
>
> HOWEVER if edit the file /ora_apps/oracle/admin/FPRD/pfile/initFPRD.ora
> and change the following parameters :-
>
> from
> db_block_buffers = 20000
> shared_pool_reserved_min_alloc = 10000
> shared_pool_reserved_size = 10000000
> shared_pool_size = 500000000
>
>
> to
> db_block_buffers = 2000
> #shared_pool_reserved_min_alloc = 10000
> #shared_pool_reserved_size = 10000000
> shared_pool_size = 5000000
>
>
> then the database starts up fine.
>
>
> The DBA's think it is a tunable paramter setting which
> needs changing.>
> We have been looking at the tunable parameters but we
> are really just taking a stab in the dark when changing
> them. Below is a list of some of the parameters
>
> Does anyone have any ideas with this problem??
>
> Thankyou in advance
>
> Ruth Reeve
> Unix System Administrator
> GCCC QLD Australia
> rreeve@goldcoast.qld.gov.au
>
>
>
>
>
>
>
> uxfmis1 (root) #> sysconfig -q ipc
> ipc:
> msg_max = 8192
> msg_mnb = 16384
> msg_mni = 1024
> msg_tql = 4096
> shm_max = 1073741824
> shm_min = 1
> shm_mni = 1024
> shm_seg = 256
> sem_mni = 1024
> sem_msl = 1024
> sem_opm = 200
> sem_ume = 200
> sem_vmx = 32767
> sem_aem = 16384
> sem_broadcast_wakeup = 1
> max_kernel_ports = 167456
> ssm_threshold = 0
> ssm_enable_core_dump = 1
> shm_allocate_striped = 1
> shm_enable_core_dump = 1
>
>
> uxfmis1 (root) #> sysconfig -q proc
> proc:
> max_proc_per_user = 2048
> max_threads_per_user = 2048
> per_proc_stack_size = 4398046511104
> max_per_proc_stack_size = 4398046511104
> per_proc_data_size = 4398046511104
> max_per_proc_data_size = 4398046511104
> max_per_proc_address_space = 4294967296
> per_proc_address_space = 4294967296
> executable_stack = 0
> autonice = 0
> autonice_time = 600
> autonice_penalty = 4
> open_max_soft = 4096
> open_max_hard = 4096
> ncallout_alloc_size = 8192
> round_robin_switch_rate = 60
> sched_min_idle = 0
> give_boost = 1
> maxusers = 2048
> num_wait_queues = 2048
> num_timeout_hash_queues = 2048
> enhanced_core_name = 0
> enhanced_core_max_versions = 16
> exec_disable_arg_limit = 0
>
>
> uxfmis1 (root) #> sysconfig -q advfs
> advfs:
> AdvfsCacheMaxPercent = 6
> AdvfsMinFragGrps = 16
> AdvfsMaxFragGrps = 48
> AdvfsAccessMaxPercent = 80
> AdvfsMinFreeAccess = 128
> AdvfsMaxFreeAccessPercent = 80
> AdvfsSyncMmapPages = 1
> AdvfsMaxDevQLen = 32
> AdvfsFavorBlockingQueue = 1
> AdvfsReadyQLim = 16384
> AdvfsDomainPanicLevel = 1
> AdvfsCacheHashSize = 0
>
>
> uxfmis1 (root) #> sysconfig -q vm
> vm:
> ubc_minpercent = 10
> ubc_maxpercent = 30
> ubc_borrowpercent = 20
> vm_max_wrpgio_kluster = 32768
> vm_max_rdpgio_kluster = 16384
> vm_cowfaults = 4
> vm_segmentation = 1
> vm_ubcpagesteal = 24
> vm_ubcfilemaxdirtypages = 4294967295
> vm_ubcdirtypercent = 40
> ubc_maxdirtywrites = 5
> ubc_maxdirtymetadata_pcnt = 70
> ubc_kluster_cnt = 32
> vm_ubcseqstartpercent = 50
> vm_ubcseqpercent = 20
> vm_csubmapsize = 1048576
> vm_ubcbuffers = 256
> vm_syncswapbuffers = 128
> vm_asyncswapbuffers = 4
> vm_clustermap = 1048576
> vm_clustersize = 65536
> vm_syswiredpercent = 80
> vm_troll_percent = 4
> vm_inswappedmin = 1
> vm_page_free_target = 1536
> vm_page_free_swap = 778
> vm_page_free_hardswap = 24576
> vm_page_free_min = 20
> vm_page_free_reserved = 10
> vm_page_free_optimal = 778
> vm_swap_eager = 1
> swapdevice = /dev/disk/dsk0b
> vm_page_prewrite_target = 3072
> vm_ffl = 1
> ubc_ffl = 1
> vm_rss_maxpercent = 100
> anon_rss_enforce = 0
> vm_rss_block_target = 778
> vm_rss_wakeup_target = 778
> kernel_stack_pages = 0
> vm_min_kernel_address = 18446741891866165248
> malloc_percpu_cache = 1
> vm_aggressive_swap = 0
> new_wire_method = 1
> vm_segment_cache_max = 50
> gh_chunks = 512
> .
> .
> .
> gh_min_seg_size = 8388608
> gh_fail_if_no_mem = 1
> private_cache_percent = 0
> gh_keep_sorted = 0
> gh_front_alloc = 1
> replicate_user_text = 1
> enable_yellow_zone = 0
> boost_pager_priority = 0
> kstack_free_target = 5
>
*********************************************************************************
This e-mail and its contents is confidential to Gold Coast City Council
and un-authorised use is strictly prohibited.
*********************************************************************************
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:48:42 EDT