Sybase NNTP forums - End Of Life (EOL)

The NNTP forums from Sybase - forums.sybase.com - are now closed.

All new questions should be directed to the appropriate forum at the SAP Community Network (SCN).

Individual products have links to the respective forums on SCN, or you can go to SCN and search for your product in the search box (upper right corner) to find your specific developer center.

forcing checkpoint more often

8 posts in Performance and Tuning Last posting was on 2011-10-13 18:40:10.0Z
Luc Van der Veurst Posted on 2011-10-07 12:25:38.0Z
From: "Luc Van der Veurst" <dba_azvub@hotmail.com>
Newsgroups: sybase.public.ase.performance+tuning
Subject: forcing checkpoint more often
Lines: 16
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5931
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
X-RFC2646: Format=Flowed; Original
X-Forwarded: by - (DeleGate/5.8.7)
X-Forwarded: by - (DeleGate/5.8.7)
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e8eefc2$1@forums-1-dub>
Date: 7 Oct 2011 05:25:38 -0700
X-Trace: forums-1-dub 1317990338 10.22.241.152 (7 Oct 2011 05:25:38 -0700)
X-Original-Trace: 7 Oct 2011 05:25:38 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12974
Article PK: 91425

Hello,

Is there in 15.5 a way to influence the frequence of the checkpoints ?

I'm bcp-ing data into a 15.5. server with trunc log on checkpoint on,
but I see long waits on log suspend.

It seems that executing checkpoint in a thresholdaction procedure wouldn't
help because a manual checkpoint only flushes dirty pages to disk but
doesn't have an impact on truncating the transaction log.

Thanks,
Luc.


Cory Sane [TeamSybase] Posted on 2011-10-07 12:33:17.0Z
From: "Cory Sane [TeamSybase]" <cory!=sane>
Newsgroups: sybase.public.ase.performance+tuning
References: <4e8eefc2$1@forums-1-dub>
In-Reply-To: <4e8eefc2$1@forums-1-dub>
Subject: Re: forcing checkpoint more often
Lines: 24
MIME-Version: 1.0
Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=response
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Windows Mail 6.0.6002.18197
X-MimeOLE: Produced By Microsoft MimeOLE V6.0.6002.18463
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e8ef18d$1@forums-1-dub>
Date: 7 Oct 2011 05:33:17 -0700
X-Trace: forums-1-dub 1317990797 10.22.241.152 (7 Oct 2011 05:33:17 -0700)
X-Original-Trace: 7 Oct 2011 05:33:17 -0700, vip152.sybase.com
X-Authenticated-User: TeamSybase
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12975
Article PK: 91424

bcp -b5000 will commit after every 5000 rows


--
Cory Sane
[TeamSybase]
Certified Sybase Associate DBA for ASE 15.0

"Luc Van der Veurst" <dba_azvub@hotmail.com> wrote in message news:4e8eefc2$1@forums-1-dub...
> Hello,
>
> Is there in 15.5 a way to influence the frequence of the checkpoints ?
>
> I'm bcp-ing data into a 15.5. server with trunc log on checkpoint on,
> but I see long waits on log suspend.
>
> It seems that executing checkpoint in a thresholdaction procedure wouldn't
> help because a manual checkpoint only flushes dirty pages to disk but
> doesn't have an impact on truncating the transaction log.
>
> Thanks,
> Luc.
>
>
>


Bret Halford Posted on 2011-10-07 15:12:05.0Z
From: Bret Halford <bret@sybase.com>
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:7.0.1) Gecko/20110929 Thunderbird/7.0.1
MIME-Version: 1.0
Newsgroups: sybase.public.ase.performance+tuning
Subject: Re: forcing checkpoint more often
References: <4e8eefc2$1@forums-1-dub>
In-Reply-To: <4e8eefc2$1@forums-1-dub>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e8f16c5$1@forums-1-dub>
Date: 7 Oct 2011 08:12:05 -0700
X-Trace: forums-1-dub 1318000325 10.22.241.152 (7 Oct 2011 08:12:05 -0700)
X-Original-Trace: 7 Oct 2011 08:12:05 -0700, vip152.sybase.com
Lines: 45
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12976
Article PK: 91429


On 10/7/2011 6:25 AM, Luc Van der Veurst wrote:
> Hello,
>
> Is there in 15.5 a way to influence the frequence of the checkpoints ?
>
> I'm bcp-ing data into a 15.5. server with trunc log on checkpoint on,
> but I see long waits on log suspend.
>
> It seems that executing checkpoint in a thresholdaction procedure wouldn't
> help because a manual checkpoint only flushes dirty pages to disk but
> doesn't have an impact on truncating the transaction log.
>
> Thanks,
> Luc.

In general the checkpoint process is woken up once a minute to scan
all the databases and checkpoint the ones that need it. To the best
of my knowledge, there is no way to change the once-a-minute timing.
For databases where truncate log on ckpt is not set, the sp_configure
parameter "recovery interval" and the amount of log activity in that
database determines whether the checkpoint process actually performs a
checkpoint in that database. For databases where truncate log is set,
a checkpoint should always occur.

On very busy systems, the checkpoint process can get overloaded and
take more than a minute to complete the checkpoints on all the
databases. If the server has more than one engine, you can configure
to have multiple checkpoint processes (up to one per engine)
using sp_configure "number of checkpoint tasks". [note: we have
received a few reports of particular databases not being checkpointed
regularly (while other databases on the same server are) when there are
multiple checkpoint tasks configured. If you observe that happening,
please open a case with tech support and ask the TSE to contact me.]

In general, if you are hitting log suspend, it sounds like your
log is undersized for the amount/size of transactions you are
running, and you should alter the database for more log space.

If you are setting up a threshold action, you can have the action
perform a "dump tran with truncate_only". You don't have to rely
on just the checkpoint to do the truncations.


Cheers,
-bret


Luc Van der Veurst Posted on 2011-10-07 17:18:07.0Z
From: "Luc Van der Veurst" <dba_azvub@hotmail.com>
Newsgroups: sybase.public.ase.performance+tuning
References: <4e8eefc2$1@forums-1-dub> <4e8f16c5$1@forums-1-dub>
Subject: Re: forcing checkpoint more often
Lines: 61
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5931
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
X-RFC2646: Format=Flowed; Response
X-Forwarded: by - (DeleGate/5.8.7)
X-Forwarded: by - (DeleGate/5.8.7)
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e8f344f$1@forums-1-dub>
Date: 7 Oct 2011 10:18:07 -0700
X-Trace: forums-1-dub 1318007887 10.22.241.152 (7 Oct 2011 10:18:07 -0700)
X-Original-Trace: 7 Oct 2011 10:18:07 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12977
Article PK: 91426

Hi Bret,

"Bret Halford" <bret@sybase.com> wrote in message
news:4e8f16c5$1@forums-1-dub...
> On 10/7/2011 6:25 AM, Luc Van der Veurst wrote:
>> Hello,
>>
>> Is there in 15.5 a way to influence the frequence of the checkpoints ?
>>
>> I'm bcp-ing data into a 15.5. server with trunc log on checkpoint on,
>> but I see long waits on log suspend.
>>
>> It seems that executing checkpoint in a thresholdaction procedure
>> wouldn't
>> help because a manual checkpoint only flushes dirty pages to disk but
>> doesn't have an impact on truncating the transaction log.
>>
>> Thanks,
>> Luc.
>
> In general the checkpoint process is woken up once a minute to scan
> all the databases and checkpoint the ones that need it. To the best
> of my knowledge, there is no way to change the once-a-minute timing.
> For databases where truncate log on ckpt is not set, the sp_configure
> parameter "recovery interval" and the amount of log activity in that
> database determines whether the checkpoint process actually performs a
> checkpoint in that database. For databases where truncate log is set,
> a checkpoint should always occur.
>
> On very busy systems, the checkpoint process can get overloaded and
> take more than a minute to complete the checkpoints on all the databases.
> If the server has more than one engine, you can configure
> to have multiple checkpoint processes (up to one per engine)
> using sp_configure "number of checkpoint tasks". [note: we have
> received a few reports of particular databases not being checkpointed
> regularly (while other databases on the same server are) when there are
> multiple checkpoint tasks configured. If you observe that happening,
> please open a case with tech support and ask the TSE to contact me.]
>
> In general, if you are hitting log suspend, it sounds like your
> log is undersized for the amount/size of transactions you are
> running, and you should alter the database for more log space.
>
> If you are setting up a threshold action, you can have the action
> perform a "dump tran with truncate_only". You don't have to rely
> on just the checkpoint to do the truncations.

Oh of course, I forgot about that, my threshold procedure already checks
for the dump tran on chkpt option and executes a dump tran with
truncate_only
:-). Thanks!

It's a new server which I'm filling up with bcp and I still had to define
the thresholds.

Luc.

PS. Cory, I'm already using a batch size of 5.000.


Luc Van der Veurst Posted on 2011-10-12 14:41:36.0Z
From: "Luc Van der Veurst" <dba_azvub@hotmail.com>
Newsgroups: sybase.public.ase.performance+tuning
References: <4e8eefc2$1@forums-1-dub> <4e8f16c5$1@forums-1-dub> <4e8f344f$1@forums-1-dub>
Subject: Re: forcing checkpoint more often
Lines: 4344
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5931
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
X-RFC2646: Format=Flowed; Response
X-Forwarded: by - (DeleGate/5.8.7)
X-Forwarded: by - (DeleGate/5.8.7)
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e95a720@forums-1-dub>
Date: 12 Oct 2011 07:41:36 -0700
X-Trace: forums-1-dub 1318430496 10.22.241.152 (12 Oct 2011 07:41:36 -0700)
X-Original-Trace: 12 Oct 2011 07:41:36 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12991
Article PK: 91441

Hello,

I'm executing bcp (sqsh built in) with batchsize 500 to populate a
table in ASE 15.5 ESD4 on Solaris Intel. The database has trunc log
on checkpoint on.

When I execute dump transaction log_db with truncate_only during
the bcp process, it takes so long that my transaction log fills up anyway,
and even when it's full, it can take minutes before it gets truncated.
(i've seen 10 minutes and more).

It's a new server, so it's very likely that I have to do some tuning..

I've attached the output of :

use log_db
go
dbcc checktable(syslogs)
go
use master
go
select getdate()
go
sp_sysmon "begin_sample"
go
dump transaction log_db with truncate_only
go
sp_sysmon "end_sample"
go
select getdate()
go
use log_db
go
dbcc checktable(syslogs)
go

Is there something in the sysmon report that indicates a problem ?

I seldom have system performance problems, so I'm not very
familiar with reading sysmon output.

Thanks,
Luc.

"Luc Van der Veurst" <dba_azvub@hotmail.com> wrote in message
news:4e8f344f$1@forums-1-dub...
> Hi Bret,
>
> "Bret Halford" <bret@sybase.com> wrote in message
> news:4e8f16c5$1@forums-1-dub...
>> On 10/7/2011 6:25 AM, Luc Van der Veurst wrote:
>>> Hello,
>>>
>>> Is there in 15.5 a way to influence the frequence of the checkpoints ?
>>>
>>> I'm bcp-ing data into a 15.5. server with trunc log on checkpoint on,
>>> but I see long waits on log suspend.
>>>
>>> It seems that executing checkpoint in a thresholdaction procedure
>>> wouldn't
>>> help because a manual checkpoint only flushes dirty pages to disk but
>>> doesn't have an impact on truncating the transaction log.
>>>
>>> Thanks,
>>> Luc.
>>
>> In general the checkpoint process is woken up once a minute to scan
>> all the databases and checkpoint the ones that need it. To the best
>> of my knowledge, there is no way to change the once-a-minute timing.
>> For databases where truncate log on ckpt is not set, the sp_configure
>> parameter "recovery interval" and the amount of log activity in that
>> database determines whether the checkpoint process actually performs a
>> checkpoint in that database. For databases where truncate log is set,
>> a checkpoint should always occur.
>>
>> On very busy systems, the checkpoint process can get overloaded and
>> take more than a minute to complete the checkpoints on all the databases.
>> If the server has more than one engine, you can configure
>> to have multiple checkpoint processes (up to one per engine)
>> using sp_configure "number of checkpoint tasks". [note: we have
>> received a few reports of particular databases not being checkpointed
>> regularly (while other databases on the same server are) when there are
>> multiple checkpoint tasks configured. If you observe that happening,
>> please open a case with tech support and ask the TSE to contact me.]
>>
>> In general, if you are hitting log suspend, it sounds like your
>> log is undersized for the amount/size of transactions you are
>> running, and you should alter the database for more log space.
>>
>> If you are setting up a threshold action, you can have the action
>> perform a "dump tran with truncate_only". You don't have to rely
>> on just the checkpoint to do the truncations.
>
> Oh of course, I forgot about that, my threshold procedure already checks
> for the dump tran on chkpt option and executes a dump tran with
> truncate_only
> :-). Thanks!
>
> It's a new server which I'm filling up with bcp and I still had to define
> the thresholds.
>
> Luc.
>
> PS. Cory, I'm already using a batch size of 5.000.
>
>
>


Bret Halford Posted on 2011-10-12 16:59:53.0Z
From: Bret Halford <bret@sybase.com>
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:7.0.1) Gecko/20110929 Thunderbird/7.0.1
MIME-Version: 1.0
Newsgroups: sybase.public.ase.performance+tuning
Subject: Re: forcing checkpoint more often
References: <4e8eefc2$1@forums-1-dub> <4e8f16c5$1@forums-1-dub> <4e8f344f$1@forums-1-dub> <4e95a720@forums-1-dub>
In-Reply-To: <4e95a720@forums-1-dub>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e95c789$1@forums-1-dub>
Date: 12 Oct 2011 09:59:53 -0700
X-Trace: forums-1-dub 1318438793 10.22.241.152 (12 Oct 2011 09:59:53 -0700)
X-Original-Trace: 12 Oct 2011 09:59:53 -0700, vip152.sybase.com
Lines: 60
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12992
Article PK: 91442

The Avg Disk I/Os returned feels low to me - I suspect
slow devices.

Disk I/O Checks
Total Disk I/O Checks 2185612.1 877732.6 990082301
n/a
Checks Returning I/O 1249838.7 501929.9 566176916
57.2 %
Avg Disk I/Os Returned n/a n/a 0.00015
n/a


You are underconfigured for "open objects"; scavanging
allocated but currently unused open objects is an expensive
operation. Don't see an obvious connection to the dump tran
running slowly, but you certainly want to bump this up
to avoid reuse.

Open Object Usage

Active n/a n/a 6499
n/a
Max Ever Used Since Boot n/a n/a 6500
n/a
Free n/a n/a 1
n/a
Reuse Requests

Succeeded n/a n/a 940
n/a


Might be helpful to get some snapshots of which code functions
the dump tran is in while it is running slowly, the function names
may provide some clues on what it is spending the most
time doing.

Open two connections, in one find the spid and issue the dump
tran

select @@spid
go
dump tran ...


In the other session, issue dbcc stacktrace(<spid>) a number of times
with a short pause in between while the dump tran is running.

dbcc traceon(3604)
go
dbcc stacktrace(<spid>)
go
...
dbcc stacktrace(<spid>)
go
[...]


-bret


Luc Van der Veurst Posted on 2011-10-12 19:21:05.0Z
From: "Luc Van der Veurst" <dba_azvub@hotmail.com>
Newsgroups: sybase.public.ase.performance+tuning
References: <4e8eefc2$1@forums-1-dub> <4e8f16c5$1@forums-1-dub> <4e8f344f$1@forums-1-dub> <4e95a720@forums-1-dub> <4e95c789$1@forums-1-dub>
Subject: Re: forcing checkpoint more often
Lines: 971
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5931
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
X-RFC2646: Format=Flowed; Response
X-Forwarded: by - (DeleGate/5.8.7)
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e95e8a1@forums-1-dub>
Date: 12 Oct 2011 12:21:05 -0700
X-Trace: forums-1-dub 1318447265 10.22.241.152 (12 Oct 2011 12:21:05 -0700)
X-Original-Trace: 12 Oct 2011 12:21:05 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12993
Article PK: 91443

Hello Bret,

Oh ok, thanks, I didn't notice the number of open object limit yet.
The limit was already higher than the production server that
has exactly the same databases and objects, does 15.5 requires
more open objects ?

I've attached the stack traces.

Storage is a Netapp filesystem, nfs, mounted with directio.

Thanks,
Luc.

"Bret Halford" <bret@sybase.com> wrote in message
news:4e95c789$1@forums-1-dub...
> The Avg Disk I/Os returned feels low to me - I suspect
> slow devices.
>
> Disk I/O Checks
> Total Disk I/O Checks 2185612.1 877732.6 990082301
> n/a
> Checks Returning I/O 1249838.7 501929.9 566176916
> 57.2 %
> Avg Disk I/Os Returned n/a n/a 0.00015
> n/a
>
>
> You are underconfigured for "open objects"; scavanging
> allocated but currently unused open objects is an expensive
> operation. Don't see an obvious connection to the dump tran
> running slowly, but you certainly want to bump this up
> to avoid reuse.
>
> Open Object Usage
>
> Active n/a n/a 6499
> n/a
> Max Ever Used Since Boot n/a n/a 6500
> n/a
> Free n/a n/a 1
> n/a
> Reuse Requests
>
> Succeeded n/a n/a 940
> n/a
>
>
> Might be helpful to get some snapshots of which code functions
> the dump tran is in while it is running slowly, the function names
> may provide some clues on what it is spending the most
> time doing.
>
> Open two connections, in one find the spid and issue the dump
> tran
>
> select @@spid
> go
> dump tran ...
>
>
> In the other session, issue dbcc stacktrace(<spid>) a number of times
> with a short pause in between while the dump tran is running.
>
> dbcc traceon(3604)
> go
> dbcc stacktrace(<spid>)
> go
> ...
> dbcc stacktrace(<spid>)
> go
> [...]
>
>
> -bret
>


Bret Halford Posted on 2011-10-13 18:40:10.0Z
From: Bret Halford <bret@sybase.com>
User-Agent: Mozilla/5.0 (Windows NT 5.1; rv:7.0.1) Gecko/20110929 Thunderbird/7.0.1
MIME-Version: 1.0
Newsgroups: sybase.public.ase.performance+tuning
Subject: Re: forcing checkpoint more often
References: <4e8eefc2$1@forums-1-dub> <4e8f16c5$1@forums-1-dub> <4e8f344f$1@forums-1-dub> <4e95a720@forums-1-dub> <4e95c789$1@forums-1-dub> <4e95e8a1@forums-1-dub>
In-Reply-To: <4e95e8a1@forums-1-dub>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4e97308a$1@forums-1-dub>
Date: 13 Oct 2011 11:40:10 -0700
X-Trace: forums-1-dub 1318531210 10.22.241.152 (13 Oct 2011 11:40:10 -0700)
X-Original-Trace: 13 Oct 2011 11:40:10 -0700, vip152.sybase.com
Lines: 109
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.performance+tuning:12994
Article PK: 91444

Hi Luc,

So it looks like it spent about a minute waiting
to get the dump lock, which probably means there was
another dump tran active on the database at the time.

And then it took about 6 minutes checkpointing the
database, which pretty much just means flushing
dirty buffers from cache to disk - so again, looks
like slow i/o.

Part of the problem could be the checkpoint process
getting overloaded, resulting in more dirty buffers
building up in the cache, resulting in more work
for checkpoint to do each time it checkpoints the
database. Something that may help that is configuring
additional checkpoint tasks - you can have up to
one per engine, and you have 4 engines.

sp_configure "number of checkpoint tasks", 4
go

But the number of i/o done over the 12 minute
sysmon sample doesn't seem all that high, so I don't
think this is the root of the issue.

-bret

On 10/12/2011 1:21 PM, Luc Van der Veurst wrote:
> Hello Bret,
>
> Oh ok, thanks, I didn't notice the number of open object limit yet.
> The limit was already higher than the production server that
> has exactly the same databases and objects, does 15.5 requires
> more open objects ?
>
> I've attached the stack traces.
>
> Storage is a Netapp filesystem, nfs, mounted with directio.
>
> Thanks,
> Luc.
>
> "Bret Halford"<bret@sybase.com> wrote in message
> news:4e95c789$1@forums-1-dub...
>> The Avg Disk I/Os returned feels low to me - I suspect
>> slow devices.
>>
>> Disk I/O Checks
>> Total Disk I/O Checks 2185612.1 877732.6 990082301
>> n/a
>> Checks Returning I/O 1249838.7 501929.9 566176916
>> 57.2 %
>> Avg Disk I/Os Returned n/a n/a 0.00015
>> n/a
>>
>>
>> You are underconfigured for "open objects"; scavanging
>> allocated but currently unused open objects is an expensive
>> operation. Don't see an obvious connection to the dump tran
>> running slowly, but you certainly want to bump this up
>> to avoid reuse.
>>
>> Open Object Usage
>>
>> Active n/a n/a 6499
>> n/a
>> Max Ever Used Since Boot n/a n/a 6500
>> n/a
>> Free n/a n/a 1
>> n/a
>> Reuse Requests
>>
>> Succeeded n/a n/a 940
>> n/a
>>
>>
>> Might be helpful to get some snapshots of which code functions
>> the dump tran is in while it is running slowly, the function names
>> may provide some clues on what it is spending the most
>> time doing.
>>
>> Open two connections, in one find the spid and issue the dump
>> tran
>>
>> select @@spid
>> go
>> dump tran ...
>>
>>
>> In the other session, issue dbcc stacktrace(<spid>) a number of times
>> with a short pause in between while the dump tran is running.
>>
>> dbcc traceon(3604)
>> go
>> dbcc stacktrace(<spid>)
>> go
>> ...
>> dbcc stacktrace(<spid>)
>> go
>> [...]
>>
>>
>> -bret
>>
>
>