Please start any new threads on our new site at https://forums.sqlteam.com. We've got lots of great SQL Server experts to answer whatever question you can come up with.

 All Forums
 SQL Server 2000 Forums
 SQL Server Administration (2000)
 PageIOLatch_SH and CXPacket

Author  Topic 

jen
Master Smack Fu Yak Hacker

4110 Posts

Posted - 2004-07-31 : 02:14:29
Hi,

I just encountered this in SQL2k and from what i read, these were bugs on SQL7.0. But since I'm running SQL2K it really doesn't make sense. The server in question has 4 cpu's and 1GB of memory.

By the way these were caused by a program that has been running before and the only difference now is that we've applied approles.

Do you think increasing the memory will solve the problem?

thanks,
Jen

derrickleggett
Pointy Haired Yak DBA

4184 Posts

Posted - 2004-08-01 : 15:08:12
I doubt it. Run a server and client side trace. See what he difference is in the record actually being produced, and the record reaching the client. CXPacket is generally related to network, although not always.

MeanOldDBA
derrickleggett@hotmail.com

When life gives you a lemon, fire the DBA.
Go to Top of Page

jen
Master Smack Fu Yak Hacker

4110 Posts

Posted - 2004-08-26 : 05:02:19
I just found out this error under system event (viewer):

The driver detected a controller error on \Device\Harddisk0.
The device, \Device\Scsi\symmpi1, is not ready for access yet.

Anyone who encountered this same errors?
Go to Top of Page

mr_mist
Grunnio

1870 Posts

Posted - 2004-08-26 : 05:25:10
Not exactly the same, but lots of pagelatch stuff with that sort of error is indicative of your hard drive subsystem either experiencing (physical) errors or being unable to cope with the demand.

-------
Moo. :)
Go to Top of Page

jen
Master Smack Fu Yak Hacker

4110 Posts

Posted - 2004-08-26 : 05:37:28
this is still connected to the topic, i issued a pathping from my workstation to the server and an ip address of 0.0.0.0 shows up and there is 100% lost of data packet from my workstation towards the 0.0.0.0 ip address and the server ip address doesn't show up at all.

what do you think?
Go to Top of Page

mr_mist
Grunnio

1870 Posts

Posted - 2004-08-26 : 05:43:28
Dodgy network card / network configuration / DNS maybe?

Is there anything in the event viewer?

-------
Moo. :)
Go to Top of Page

jen
Master Smack Fu Yak Hacker

4110 Posts

Posted - 2004-08-26 : 05:48:52
Nope, only those two errors keep coming up consistently

the odd thing is the 0.0.0.0 ip address destination, it should have returned the destination ip address i specified on the pathping.

something must have gone wrong between transfer?
Go to Top of Page

jen
Master Smack Fu Yak Hacker

4110 Posts

Posted - 2004-08-26 : 06:33:08
Hi All,

For those interested here's what i found out:

Data sent is lost from workstation to server due to firewall

A month ago, our techsupport decided to implement mini-firewalls, and it so happens that after this, unusual problems came up. Timeout expired was a daily reminder that something is really wrong.

They kept telling us that if we can ping the servers then there's no problem with network even if we encounter connection timeout and general network failures. I didn't believe them and still don't coz log files don't lie right? Heck, they can't even troubleshoot the problem.

Well, the moral of the story? Don't believe what they always say, even if it's out of your scope if it's killing you then do something to prove your cause.

Thanks for the replies and support...
Go to Top of Page

mr_mist
Grunnio

1870 Posts

Posted - 2004-08-26 : 07:22:38
Firewalls in between the server and client were the cause of many problems for me in the past. If it's not general network errors then it's disconnected sessions after 24hrs.

-------
Moo. :)
Go to Top of Page

jen
Master Smack Fu Yak Hacker

4110 Posts

Posted - 2004-08-26 : 08:41:46
Actually Moo,

I had a blast with this one. When I first came in, the netadmins were all doing a good job of being superior and I-know-it-all-attitude.

I informed them, there's a virus trying to infect my machine (had an audit log). They didn't believe me. You know what? That Virus just made my day!

It so happens that the firewall they'd set up, isolated the virus on a part of the network and it just caused so much congestion that the latches and cxpackets were produced.

And to top it all, it's a physical firewall, it's not even a VPN.So they're gonna have a field day just isolating that virus.

Good luck to them, we're more than 2,000 machines and these machines should not be offlined for more than 30 minutes.

I know I'm being bad, but hey just this once
Go to Top of Page
   

- Advertisement -