Re: IDS w/ ADS 8.21 : Startup problem.

---------- From: David Williams [SMTP:djw@smooth1.demon.co.uk] Sent: Saturday, September 25, 1999 3:21 AM To: informix-list@iiug.org Subject: Re: IDS w/ ADS 8.21 : Startup problem. In article <7sfv9p$719$1@news.xmission.com>, Maria Seagull <maria_seagull@yahoo.com> writes > >Hi, >Here are the details of the problem we have , your >attention and help is highly appreciated. Thanks. >- Maria > >System : >Informix server : IDS with ADS Version 8.21.UD4D1 >Unix OS : HP release B.11.00 HP9000/V2200 > >Context: We were finetuning the engine shared memory, >when we faced problems >bringing up the server with a certain set of onconfig >values(detailed below). > >Problem: >Bringing up the server on-line with the following >onconfig parameters: >BUFFERS 60000 # Max. no. of shared buffers (4kb >pages) >SHMVIRTSIZE 345000 # initial virtual shared memory>segment size > >We were earlier able to bring it up with the following >parameters: >BUFFERS 50000 # Max. no. of shared buffers (4kb >pages) >SHMVIRTSIZE 305000 # initial virtual shared memory>segment size > >So in short we took some from virtual sh. mem and gave >some to buffers >(or so we thought ..:-) ) but the engine didn't seem >to like it and failed to come >up . > What errors do you get? Out of shared memory, check system SHMALL. Mt_shm_init : can't create virtual segment. Check kernel parameters allow enough shared memory segments to be created. I think we have kept it as per the release notes. Could you please explain which parameters to check and what is the usualy recommended/optimum values. TIA. Perhaps it needs that much virtual shared memory to start? Well it starts with lower SHMVIRTSIZE , so we cannot say that it could need that much virtual sh. Mem. But the question is why should there be a problem when we Are merely giving some memory from virtual sh. Mem. To buffers. And why is there no problem , if we decrease the SHMVIRTSIZE to say 180000. start using oninit -v?? We used xctl 'C oninit 'y. What does the 'v option do?TIA. Thanks David. - Maria > >Question: Why did it fail to come up? How to bring it >up with increased buffers? >Any ideas,hints,suggestions are welcome >and would be appreciated. TIA. > > > > >Relevant Details from the onconfig : >(In case u need more details to "analyse this" , just >let me know. Thanks ) > > > ># Shared Memory Parameters on onconfig file > > >LOCKS 200000 # Maximum number oflocks ># this does not work BUFFERS 60000 # >Maximum number of shared buffers (4 kb pages) >BUFFERS 50000 # Maximum number ofshared >buffers (4 kb pages) > >NUMAIOVPS 2 # Number of IO vps>NUMFIFOVPS 2 # Number of FIFO vps >#PHYSBUFF 32 # Physical log buffer size (Kbytes) >#LOGBUFF 32 # Logical log buffer size (Kbytes) >PHYSBUFF 64 # Physical log buffer size (Kbytes) >LOGBUFF 64 # Logical log buffer size (Kbytes) >LOGSMAX 110 # Maximumnumber of logical log >files >CLEANERS 32 # Number of buffer>cleaner processes >#CLEANERS 64 # Number of buffer >cleaner processes >SHMBASE 0x0 # Shared memory baseaddress >#SHMBASE 0 # Shared memory base address > > >SHMVIRTSIZE 250000 # initial virtualshared >memory segment size >#SHMVIRTSIZE 345000 # initial virtual shared >memory segment size > >#SHMADD 16200 # Size of new shared >memory segments (Kbytes) ># cberner 09/11/1999 >SHMADD 500 # Size of new shared>memory segments (Kbytes) >#SHMADD 16200 # Size of new shared >memory segments (Kbytes) >SHMTOTAL 0 # Total shared memory >(Kbytes). 0=>unlimited >#SHMTOTAL 1600000 # Total shared >memory (Kbytes). 0=>unlimited >CKPTINTVL 600 # Check pointinterval >(in sec) >LRUS 24 # Number of LRU queues>#LRUS 32 # Number of LRU queues >LRU_MAX_DIRTY 20 # LRU percent dirtybegin cleaning >limit >LRU_MIN_DIRTY 15 # LRU percent dirtyend cleaning >limit >LTXHWM 50 # Long transactionhigh water mark >percentage >LTXEHWM 60 # Longtransaction high water mark >(exclusive) >TXTIMEOUT 300 # Transaction timeout (in sec) >STACKSIZE 32 # Stack size (Kbytes)> > >System Configuration > >SERVERNUM 1 # Unique idcorresponding to a IDS AD >instance > > >NETTYPE ipcshm,2,100,CPU #Override sqlhosts >nettype parameters >#NETTYPE ipcshm,2,100,NET # Override sqlhosts >nettype parameters >NETTYPE soctcp,2,200,NET #Override sqlhosts >nettype parameters >DEADLOCK_TIMEOUT 60 # Max time to wait oflock in >distributed env. >RESIDENT 1 # Forced residency flag (Yes = 1, No = 0) > >MULTIPROCESSOR 1 # 0 for>single-processor, 1 for multi-processor >NUMCPUVPS 2 # Number of user (cpu)vps >SINGLE_CPU_VP 0 # If non-zero, limit>number of cpu vps to one > >NOAGE 0 # Process aging >AFF_SPROC 0 # Affinity startprocessor >AFF_NPROCS 0 # Affinity number ofprocessors > > > > > DBSPACETEMP: ># IDS AD equivalent of DBTEMP for SE. This is the list >of dbspaces ># that the IDS AD SQL Engine will use to create temp >tables etc. ># The list can be specified using various preset >"sets" that are ># ALL - all dbspaces can be used for temp files ># TEMP - use only temporary dbspaces for temp files ># NOTCRITICAL - use only non-critical dbspaces for >temp files ># Critical dbspaces are the rootdbspaces and >dbspaces containing ># log files. ># The list can also be specified as an explicit list >of dbspaces/dbslices ># that exist when the IDS AD system is brought online. > If not specified, ># or if all dbspaces specified are invalid, various ad >hoc queries will create ># temporary files in /tmp instead. If a dbspaces in >this list is ># added to the system at a later point in time, it is >automatically added ># to the list of dbspaces that can be used for temp >purposes. > >DBSPACETEMP NOTCRITICAL # Default tempdbspaces > >__________________________________________________ >Do You Yahoo!? >Bid and sell for free at http://auctions.yahoo.com -- David Williams _________________________________________