Could not spawn process_loginread thread
-
I have come across a SQL Server 2000 Enterprise Edition machine that periodically starts logging 'SQL Server could not spawn process_loginread thread.' for a while before it finally goes down and then restarts automatically. I have not been able to find either a cause for this or a solution, so now I'm asking you guys if you've seen this on any of your servers? As I said, the server is a SQL2K Ent Ed SP2, Win2K Standard Server SP2. It's got 4 GB of memory and has AWE enabled in order to use this, but has max server memory set to 2500 since there are a couple of other smaller applications running on the server.
Periodically it starts logging said message, and does so for about a day or so before the server finally goes down. I'm not 100% sure, but it also seems to occur more often in the end. At least it's not occurring at a specific interval. We have not been able to find any special action or situation in the application that could start this. The server doesn't seem to be bogged down by this though, until it goes down it works more-or-less as usual, no exact measuring done though.
In the knowledge base I've found two articles mentioning this error message, but both are supposedly corrected in SP1, and both of them are connected with functionality not used on the server (sp_OA-procs and stack dumps). So, has anyone else ever seen this?
Chris Hedgate @ Apptus Technologies (http://www.apptus.se)
-
No but take a look here http://webferret.search.com/click?wf,Server+could+not+spawn+process_loginread+thread.,,support.microsoft.com%2Fsupport%2Fkb%2Farticles%2Fq282%2F2%2F29.ASP,,fastsearch . This sounds like the issue especially since you point out the sp_OA-procs relation.
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
-
That's one of the articles I mentioned, and it is supposedly corrected in SP1 plus it discusses problems with sp_OA-procs, which are not used at all on the server. So I'm not too sure that this is the problem. Thanks for the reply though.
--
Chris Hedgate @ Apptus Technologies (http://www.apptus.se)
-
Ha ha, sorry about that I was reading crossed eyed apparently as I missed that section. However you may still have dumps going on that can explain what is happening. Check the log folder under the SQL install (the same place the errorlogs and agent logs are located) and see if any .dmp files exist.
Also found this in another thread off site and might help.
quote:
When the server is allowing max server memory unlimited, and the server isset
to dynamic memory allocation, then these error should only showing up in a
short period of time when the memory is tight for the sql server to add
more
connections, it would need some time to release some of the allocated
memory,
likely from the buffer pool to adjust the memory request, then the error
will
go a way after few seconds.
If the problem persist for a long period of time, no one can connect, then
check to see if the sql server has been set to a fix memory configuration.
In
that case, the sql server min and max server memory were set to the same
value
and doesn't allow to release any of the memory the sql server has
allocated,
and there aren't enough memory to allocate additional connection then only
a recycle of the sql server would help to get out of the
problem.
Run dbcc freeproccache may help to release some of the needed memory.
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
-
Thanks again Antares. I checked the server min setting, but it was set to 1500, with max = 2500. So that doesn't seem to be it. Next stop is checking for dumpfiles....
--
Chris Hedgate @ Apptus Technologies (http://www.apptus.se)
-
Hmm... there was actually some dump files, but they where much older than when this first started occurring, so that shouldn't be it either. Just to be sure I did as it said in the workaround of the kb describing this, renaming dbhelp.dll and sqlimage.dll. Problem is it can take days until next crash....
--
Chris Hedgate @ Apptus Technologies (http://www.apptus.se)
-
Well, I spend a lot of time on the internet and use various tools for research but if you need to look around a bit for possible related items take a look at http://www.webferret.com which works as a great web search tool. Otherwise, if I come across anything I will pass on, and let us know what happens.
Just a thought if it does start happening again and you catch before the server stops try the suggestion:
Run dbcc freeproccache may help to release some of the needed memory.
To see if it is a memory issue. Just keep in mind it may slow processing of queries (but if the server stops that will happen anyway, so will be worth a try).
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
Edited by - antares686 on 10/08/2002 6:40:43 PM
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply