SQL Anywhere database - ASA 7.0 database freezing

Author

ASA 7.0 database freezing

nirupa

2005-05-10, 3:23 am

hi

we have an ASA 7.0 database around 31/2 gb in size running on win 2k advanced
server. there are 66 clients conencting to this from same and different subnet.
i have set the cache size to around 900 mb for this. it is a primavera
expediiton application.
frequently we find the database freezing. at this time it does not allow new
connections and existing connections cannot work further.
but when i check in the server, the service is running. so waht i do is stop and
restart the service, then the clients can connect etc.

any clue on why this is happening ?
---== Posted via the PFCGuide Web Newsreader ==---
http://www.mcse.ms/_newsgroups/group_list.asp

Nick Elson

2005-05-10, 9:23 am

4 possibilities come to mind.

1 - Since you are running with Win2K make sure you have
a fairly recent level of 7. Upgrading to Win2K introduced
a number of new problems that were not fixed until late in
7.0.3 and early 7.0.4. I would test with the latest 7.0.4
ebf.

2 - The server may just be no longer able to service the network.
Attempting to connect locally, on the same machine as the
server, connects over shared memory and should always
be possible. If that works, then you have a network
related issue. (A side note to this is some of the W2K
issues, see note 1 above, were network related.)

3 - You could just be blocking in the application. If this 'hang'
is due to normal record locking running both:

call sa_conn_info( )
and
call sa_locks( )

should give you more insight into the problem. Of course
you need to already be connected to do so. Leave a
dbisql session logged in from before you hit this problem.
Launching that on the server machine, and starting it ahead
of time should bypass any of the issues normally affecting
these kinds of problems. Of course you do need to plan
ahead of time for this to work.

or

4 - You could just be blocking on a resource. That could just
be due to:

- too many threads being used up (see -gn and consider
bumping it higher than the default to see if that helps any)

- your OS and/or hardware could just be exhausted
look for CPU, activity on the disk, and disk channel
capacity getting exhausted

Of course, all of the above assumes that the ASA serer is running
normally and correctly. I guess the server could just be asserting.
Is there a -o switch on that server? Is there an entry in that server
log file or in the system event log saying anything about this? If there
is no indication (either way) try running the server in the foreground
to see if the console shows you anything more.

Good luck

"nirupa" < radhika_sridhar@hotm
ail.com> wrote in message
news:42806e33$1@foru
ms-2-dub...
> hi
>
> we have an ASA 7.0 database around 31/2 gb in size running on win 2k
> advanced
> server. there are 66 clients conencting to this from same and different
> subnet.
> i have set the cache size to around 900 mb for this. it is a primavera
> expediiton application.
> frequently we find the database freezing. at this time it does not allow
> new
> connections and existing connections cannot work further.
> but when i check in the server, the service is running. so waht i do is
> stop and
> restart the service, then the clients can connect etc.
>
> any clue on why this is happening ?
> ---== Posted via the PFCGuide Web Newsreader ==---
> http://www.mcse.ms/_newsgroups/group_list.asp

nirupa

2005-05-11, 3:23 am

hi help

some more info.
my ASA version is 7.0.3.
when the system freezes , i checked the memory and cpu utilisation and found it
to be very very less. so that cannot be the problem.also the server has 4 gb
RAM.

but one thing i noticed today in performance monitor is
Requests : Active : 428
Unshceduled : 0

My setting of -gn in the parameter file is 50. we have 66 client licenses. that
means maximum of 66 clients can connect at a time.

i read the ASA refernce manual wherein they have mentioned -gn cannot be greater
than the no. of server connections and the default for network server is 20.he
also mentions that if the total of (active+unscheduled)
> -gn then you might
increase the -gn value.

now i am getting a doubt ,should i increase

-gn to 66 (no. of maximum client connections possible at one time )
or

-gn to 428 which is the total of (active+unscheduled )
but as per the manual , this no. cannot be greater than the no. of server
connections. then why is it showing 428 instead of 66 ?

Pls. clarify on this

regards

On 10 May 2005 06:26:49 -0700,
in sybase.public.sqlanywhere.general
Nick Elson < no_spam_nicelson@syb
ase.com> wrote:
>4 possibilities come to mind.
>
>1 - Since you are running with Win2K make sure you have
> a fairly recent level of 7. Upgrading to Win2K introduced
> a number of new problems that were not fixed until late in
> 7.0.3 and early 7.0.4. I would test with the latest 7.0.4
> ebf.
>
>2 - The server may just be no longer able to service the network.
> Attempting to connect locally, on the same machine as the
> server, connects over shared memory and should always
> be possible. If that works, then you have a network
> related issue. (A side note to this is some of the W2K
> issues, see note 1 above, were network related.)
>
>3 - You could just be blocking in the application. If this 'hang'
> is due to normal record locking running both:
>
> call sa_conn_info( )
> and
> call sa_locks( )
>
> should give you more insight into the problem. Of course
> you need to already be connected to do so. Leave a
> dbisql session logged in from before you hit this problem.
> Launching that on the server machine, and starting it ahead
> of time should bypass any of the issues normally affecting
> these kinds of problems. Of course you do need to plan
> ahead of time for this to work.
>
>or
>
>4 - You could just be blocking on a resource. That could just
> be due to:
>
> - too many threads being used up (see -gn and consider
> bumping it higher than the default to see if that helps any)
>
> - your OS and/or hardware could just be exhausted
> look for CPU, activity on the disk, and disk channel
> capacity getting exhausted
>
>
>Of course, all of the above assumes that the ASA serer is running
>normally and correctly. I guess the server could just be asserting.
>Is there a -o switch on that server? Is there an entry in that server
>log file or in the system event log saying anything about this? If there
>is no indication (either way) try running the server in the foreground
>to see if the console shows you anything more.
>
>
>Good luck
>
>
>"nirupa" < radhika_sridhar@hotm
ail.com> wrote in message
> news:42806e33$1@foru
ms-2-dub...
>
>

nirupa

2005-05-11, 3:23 am

hi nick
i just missed some points.

you have mentioned about EBF. where can i get this.
you have also mentioned about running the server in foreground. how do i do
this

your quick help is appreciated

regards

On 10 May 2005 23:15:50 -0700,
in sybase.public.sqlanywhere.general
nirupa < radhikajbr@dubaiinte
rnetcity.ae> wrote:
>hi help
>
>some more info.
>my ASA version is 7.0.3.
>when the system freezes , i checked the memory and cpu utilisation and found it
>to be very very less. so that cannot be the problem.also the server has 4 gb
>RAM.
>
>but one thing i noticed today in performance monitor is
>Requests : Active : 428
> Unshceduled : 0
>
>My setting of -gn in the parameter file is 50. we have 66 client licenses. that
>means maximum of 66 clients can connect at a time.
>
>i read the ASA refernce manual wherein they have mentioned -gn cannot be
greater
>than the no. of server connections and the default for network server is 20.he
>also mentions that if the total of (active+unscheduled)
> -gn then you might
>increase the -gn value.
>
>now i am getting a doubt ,should i increase
>
> -gn to 66 (no. of maximum client connections possible at one time )
> or
>
> -gn to 428 which is the total of (active+unscheduled )
> but as per the manual , this no. cannot be greater than the no. of
server[color=darkred
]
>connections. then why is it showing 428 instead of 66 ?
>
>Pls. clarify on this
>
>regards
>
>On 10 May 2005 06:26:49 -0700,
> in sybase.public.sqlanywhere.general
>Nick Elson < no_spam_nicelson@syb
ase.com> wrote:

Nick Elson

2005-05-11, 11:23 am

I'd give -gn 100 a try and adjust up or down depending upon the results.

"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
news:4281a316$1@foru
ms-1-dub...[color=darkred]
> hi help
>
> some more info.
> my ASA version is 7.0.3.
> when the system freezes , i checked the memory and cpu utilisation and
> found it
> to be very very less. so that cannot be the problem.also the server has 4
> gb
> RAM.
>
> but one thing i noticed today in performance monitor is
> Requests : Active : 428
> Unshceduled : 0
>
> My setting of -gn in the parameter file is 50. we have 66 client licenses.
> that
> means maximum of 66 clients can connect at a time.
>
> i read the ASA refernce manual wherein they have mentioned -gn cannot be
> greater
> than the no. of server connections and the default for network server is
> 20.he
> also mentions that if the total of (active+unscheduled)
> -gn then you
> might
> increase the -gn value.
>
> now i am getting a doubt ,should i increase
>
> -gn to 66 (no. of maximum client connections possible at one time )
> or
>
> -gn to 428 which is the total of (active+unscheduled )
> but as per the manual , this no. cannot be greater than the no. of
> server
> connections. then why is it showing 428 instead of 66 ?
>
> Pls. clarify on this
>
> regards
>
> On 10 May 2005 06:26:49 -0700,
> in sybase.public.sqlanywhere.general
> Nick Elson < no_spam_nicelson@syb
ase.com> wrote:

Nick Elson

2005-05-11, 11:23 am

Since 7.0.x has reached the end of it's active support life
you should go right to 7.0.4 and apply the latest ebf for 7.0.4.

You do so by going to (and logging into)

http://downloads.sybase.com/swd/summary.do

and once to get to the SQL Anywhere Studio page, change
the drop down box that currently shows

"in the last 6 months"
to

"in all months"

and then click the [Go!] button to the right of that.

Once that refreshes scroll down to the section

"Platform: Windows x86"

and from there download these two picks

SQL Anywhere Studio - Maintenance Release (English), Upgrade of 7.0.x
to 7.0.4

and

SQL Anywhere Studio - Express Bug Fix, Update of 7.0.4 to build 3541

[and the read-me documentation that are the icons just to the left of each
of those]

Apply the Maintenance release first. That takes you to 7.0.4. Then
apply the ebf last.

"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
news:4281b2e7$1@foru
ms-1-dub...[color=darkred]
> hi nick
> i just missed some points.
>
> you have mentioned about EBF. where can i get this.
> you have also mentioned about running the server in foreground. how do i
> do
> this
>
> your quick help is appreciated
>
> regards
>
> On 10 May 2005 23:15:50 -0700,
> in sybase.public.sqlanywhere.general
> nirupa < radhikajbr@dubaiinte
rnetcity.ae> wrote:
> greater
> server

nirupa

2005-05-12, 3:24 am

Hi nick

thanks for the response. i have few things that is bothering me as i am not able
to conclude at all.

1. will increaseing the -gn value have an effect in the performance ?
is there any formula or technique for arriving a combination of -gn and
-c,-ch parameter

2. will applying ebf solve this issue ?
3. you have mentioned about some unresolved win 2k network issues with prior
versions of 7.0.3. where can i get more info on this?

4. how do i run the dbf in the foreground ?

one more detail i want to share with you.
generally if i want to stop the database, i go to sybase central ,right click
the service and stop it. then i get a message saying 'client connections are
still open do you want to close anyway?

but when the system freezes, and i try to stop the service the same way it stops
without any message. that means it has automatically disconnected all the
clients ? am i correct ? this also means the database is running but not
allowing any new client connections ? am i correct ?
but when i stop and start the service again, it allows client connections ? in
this case how can it be a network issue ? if it was a network issue after
stopping and starting also it should not allow clients ?

pls. clarify this point which is bothering me much as i want to know if the
problem is in the server setup or network. i want to isolate this first.

also so far there is no -o in the server, i shall give this and post you on any
messages .

pls. note that when this happens , there is no message recorded in the windows
event viewer.

regards

On 11 May 2005 09:28:23 -0700,
in sybase.public.sqlanywhere.general
Nick Elson < no_spam_nicelson@syb
ase.com> wrote:
>Since 7.0.x has reached the end of it's active support life
>you should go right to 7.0.4 and apply the latest ebf for 7.0.4.
>
>You do so by going to (and logging into)
>
> http://downloads.sybase.com/swd/summary.do
>
>and once to get to the SQL Anywhere Studio page, change
>the drop down box that currently shows
>
> "in the last 6 months"
>to
>
> "in all months"
>
>and then click the [Go!] button to the right of that.
>
>Once that refreshes scroll down to the section
>
> "Platform: Windows x86"
>
>and from there download these two picks
>
> SQL Anywhere Studio - Maintenance Release (English), Upgrade of 7.0.x
>to 7.0.4
>
>and
>
> SQL Anywhere Studio - Express Bug Fix, Update of 7.0.4 to build 3541
>
>[and the read-me documentation that are the icons just to the left of each
>of those]
>
>Apply the Maintenance release first. That takes you to 7.0.4. Then
>apply the ebf last.
>
>
>"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
> news:4281b2e7$1@foru
ms-1-dub...
>
>

Nick Elson

2005-05-12, 9:23 am

[color=darkred]

Only way to know is to test it out, in your setting, to see how it
affects you. It could just improve performance for you and if
it does, then the answer is YES.
[color=darkred]

If we know what the core problem really was then maybe I
could answer that. But check out the read me files that come
with those fixes to see for yourself how much and what has
been fixed already.
[color=darkred]

see the point about the read me file
[color=darkred]

Don't run it as a service. Create a .BAT file or short cut. Type it
into a CMD window. .... your choice.
[color=darkred]

That probably means it has crashed with some sort of
FATAL ERROR or ASSERTION ERROR. Find out
which and you may know what needs to happen next.
Add your -o switch to find out.

"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
news:4282ea36@forums
-1-dub...[color=darkred]
> Hi nick
>
> thanks for the response. i have few things that is bothering me as i am
> not able
> to conclude at all.
>
> 1. will increaseing the -gn value have an effect in the performance ?
> is there any formula or technique for arriving a combination of -gn and
> -c,-ch parameter
>
> 2. will applying ebf solve this issue ?
> 3. you have mentioned about some unresolved win 2k network issues with
> prior
> versions of 7.0.3. where can i get more info on this?
>
> 4. how do i run the dbf in the foreground ?
>
> one more detail i want to share with you.
> generally if i want to stop the database, i go to sybase central ,right
> click
> the service and stop it. then i get a message saying 'client connections
> are
> still open do you want to close anyway?
>
> but when the system freezes, and i try to stop the service the same way it
> stops
> without any message. that means it has automatically disconnected all the
> clients ? am i correct ? this also means the database is running but not
> allowing any new client connections ? am i correct ?
> but when i stop and start the service again, it allows client connections
> ? in
> this case how can it be a network issue ? if it was a network issue after
> stopping and starting also it should not allow clients ?
>
> pls. clarify this point which is bothering me much as i want to know if
> the
> problem is in the server setup or network. i want to isolate this first.
>
> also so far there is no -o in the server, i shall give this and post you
> on any
> messages .
>
> pls. note that when this happens , there is no message recorded in the
> windows
> event viewer.
>
> regards
>
>
>
>
>
> On 11 May 2005 09:28:23 -0700,
> in sybase.public.sqlanywhere.general
> Nick Elson < no_spam_nicelson@syb
ase.com> wrote:

nirupa

2005-05-13, 11:23 am

hi nick

thanks for the response. i added the -o switch and i find the following error
message recorded whenever the system freezes

TCP listener shutting down(64)

your help is very much appreciated.

regards

On 12 May 2005 06:05:18 -0700,
in sybase.public.sqlanywhere.general
Nick Elson < no_spam_nicelson@syb
ase.com> wrote:
>
>Only way to know is to test it out, in your setting, to see how it
>affects you. It could just improve performance for you and if
>it does, then the answer is YES.
>
>
>If we know what the core problem really was then maybe I
>could answer that. But check out the read me files that come
>with those fixes to see for yourself how much and what has
>been fixed already.
>
>
>see the point about the read me file
>
>
>Don't run it as a service. Create a .BAT file or short cut. Type it
>into a CMD window. .... your choice.
>
>
>That probably means it has crashed with some sort of
>FATAL ERROR or ASSERTION ERROR. Find out
>which and you may know what needs to happen next.
>Add your -o switch to find out.
>
>"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
> news:4282ea36@forums
-1-dub...
>
>

Chris Keating $iAnywhere Solutions$

2005-05-13, 11:23 am

This error means "The specified network name is no longer available.". On
the surface, it looks like the network is flakey. Have you looked at your
network stablility?

"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
news:4284c55d$1@foru
ms-1-dub...[color=darkred]
> hi nick
>
> thanks for the response. i added the -o switch and i find the following
> error
> message recorded whenever the system freezes
>
> TCP listener shutting down(64)
>
> your help is very much appreciated.
>
> regards
>
>
>
> On 12 May 2005 06:05:18 -0700,
> in sybase.public.sqlanywhere.general
> Nick Elson < no_spam_nicelson@syb
ase.com> wrote:

nirupa

2005-05-14, 1:23 pm

hi help
i dont think there is network instability. becoz exactly at this point of time
when this error occurs, all clients are able to access shared files in the same
directory where the ASA database runs.
as the error message says TCP listener is shutting down(64)
i am having a doubt.

1. is there any limit for the no. of connections for the TCP listener service.
2. is the listener service running separte of database service as in case of
oracle or are they both one and the same.
3. is there a parameter file that lists some details of this service
4. or lastly can the listener also stop if the dbf is running short of resouces
like memory
5. is there any formula for RAM requirement for ASA database ?
for a databsae of size 3 gb,how much RAM should i allocate?

On 13 May 2005 08:32:20 -0700,
in sybase.public.sqlanywhere.general
Chris Keating $iAnywhere Solutions$ < Spam_NoThanks_keatin
g@iAnywhere.com>
wrote:
>This error means "The specified network name is no longer available.". On
>the surface, it looks like the network is flakey. Have you looked at your
>network stablility?
>
>
>"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
> news:4284c55d$1@foru
ms-1-dub...
>
>

nirupa

2005-05-16, 3:23 am

hi help

i found this detail in the sybase site.

Solved Cases (#41093541)-SQL Anywhere Studio 0600-PC (Solved Cases)
Summary: Win32 server may stop listening for TCP connections if socket errors
occur Versions affected: 6.x, 7.x, 8.x Description: When a new connection is
made to the server, the message 'TCP Listener shutting down (2?)' could be
displayed, after
Category: Support and Trouble Shooting Date: Mar 18, 2002

i tried getting this article, but i couldnt. can you pls. get me this article
here.

thanks in advance.

waiting eagerly for your reply.

On 13 May 2005 08:32:20 -0700,
in sybase.public.sqlanywhere.general
Chris Keating $iAnywhere Solutions$ < Spam_NoThanks_keatin
g@iAnywhere.com>
wrote:
>This error means "The specified network name is no longer available.". On
>the surface, it looks like the network is flakey. Have you looked at your
>network stablility?
>
>
>"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
> news:4284c55d$1@foru
ms-1-dub...
>
>

nirupa

2005-05-17, 3:23 am

Hi nick

Pls. reply. I am waiting for your reply.

On 14 May 2005 10:43:09 -0700,
in sybase.public.sqlanywhere.general
nirupa < radhikajbr@dubaiinte
rnetcity.ae> wrote:[color=darkred
]
>hi help
>i dont think there is network instability. becoz exactly at this point of time
>when this error occurs, all clients are able to access shared files in the same
>directory where the ASA database runs.
>as the error message says TCP listener is shutting down(64)
>i am having a doubt.
>
>1. is there any limit for the no. of connections for the TCP listener service.
>2. is the listener service running separte of database service as in case of
>oracle or are they both one and the same.
>3. is there a parameter file that lists some details of this service
>4. or lastly can the listener also stop if the dbf is running short of resouces
>like memory
>5. is there any formula for RAM requirement for ASA database ?
>for a databsae of size 3 gb,how much RAM should i allocate?
>
>
>
>
>
>
>On 13 May 2005 08:32:20 -0700,
> in sybase.public.sqlanywhere.general
>Chris Keating $iAnywhere Solutions$ < Spam_NoThanks_keatin
g@iAnywhere.com>
>wrote:

Nick Elson

2005-05-17, 9:23 am

I should first point out that this is *not* a helpdesk. People volunteer
their time and information to these newsgroups as a way to share
information amongst our friends and customers. Occasionally,
iAnywhere personnel may drop in from time to time to offer a couple
of ideas as well but there is no guarantees that will happen.

If you are in need of support urgently, you should contact the support
teams that cover your region them directly; see

http://www.ianywhere.com/support/co...ct_support.html

As to your 'hit' another self-help tip, would be to search for these
things in the readme for the 7.0.4 Maintenance Release and last
EBFs for 7.0.4. Trying to get access to other persons support
case files will likely fail. If you have a support plan you can always
look at your own case files but not others. While we attempt to
disseminate technical information from such cases we must not
expose customer sensitive information and so information from
support case files tend not to be generally available.

As to the specific bug fix mentioned in the text of that case, that
would be this one:

QTS# 238300
Versions affected: 6.x, 7.x
Versions fixed: 6.0.4EBF (3730), 7.0.2EBF (1483), 7.0.3, 8.0
Modules affected: dbserv6.dll, dbserv7.dll, dbserv7.dll

Description: When a new connection is made to the server, the message
"TCP Listener shutting down (2:10038)" could be displayed, after which,
no new TCP connections will be accepted by the server.

So if you are already at 7.0.3GA (i.e. no ebfs) or higher you already have
this fix. Besides, you are getting error 64 and not 10038 so this is not
your issue or required fix.

To prove to yourself what this error means do the following yourself

On any Windows 32bit OS (NT4.0 or higher) submit the following
command and note that the text is coming back from the Microsoft
OS

C:\>net helpmsg 64
"The specified network name is no longer available. "

How you are hitting this is unknown but it *IS* a network error
coming back the DNS-CLIENT layer.

Turning on the -Z on the server command line might tell you
a little more about when this is happening but otherwise,
you probably need to get a network sniffer into this equation
and work with your IT personnel or your outsourced support
company.

"nirupa" < radhikajbr@dubaiinte
rnetcity.ae> wrote in message
news:42897d09$1@foru
ms-1-dub...[color=darkred]
> Hi nick
>
> Pls. reply. I am waiting for your reply.
>
>
>
> On 14 May 2005 10:43:09 -0700,
> in sybase.public.sqlanywhere.general
> nirupa < radhikajbr@dubaiinte
rnetcity.ae> wrote:

Manney

2005-05-19, 3:23 am

I too am having a very similar problem, and have just posted the group for
advice ( that was before I saw this thread, sorry) but with a far smaller DB
file 115Mb running ASA 7.0.4 on a win2003 server.

regards

Manney

nirupa

2005-05-24, 7:23 am

did you get any response ?

On 19 May 2005 01:12:12 -0700,
in sybase.public.sqlanywhere.general
Manney <nospam@mycompany.com> wrote:
>I too am having a very similar problem, and have just posted the group for
>advice ( that was before I saw this thread, sorry) but with a far smaller DB
>file 115Mb running ASA 7.0.4 on a win2003 server.
>
>regards
>
>Manney
>
>

Sponsored Links