View Full Version : Networking issue?
We've recently upgraded most of our switches at my work with Netgear 24 port gigabit switches.
We've also had an increrase in problems with users mailboxes (outlook .pst files) getting corupted.
Is there any way to tell if this is related to the newer switches? Is there a chance they might be more prone to errors?
You're putting PSTs on network drives?
You're putting PSTs on network drives?
No. But with roaming profiles, they are copied over to the profile server share when the user logs out. We have some pretty big .pst files. We probably aren't going to be able to enforce anything but the 2GB hard limit on them until after we get in exchange.
Exchange fixes all :)
In fact, it should be your priority to get it in place..
Hmmm...I can't really see how the new switch could be related then. It's possible packet loss is causing them to be copied over in a corrupted state, but I would imagine you would see similar symptoms in other files as well (unless nothing else comes close to those PSTs). PSTs are typically pretty fragile, so it's possible. Are these managed switches, or just plug-n-go?
Exchange fixes all :)
In fact, it should be your priority to get it in place..
Honest, you don't have to convince me. It is on the way, it just takes a while to process things here.
Hmmm...I can't really see how the new switch could be related then. It's possible packet loss is causing them to be copied over in a corrupted state, but I would imagine you would see similar symptoms in other files as well (unless nothing else comes close to those PSTs). PSTs are typically pretty fragile, so it's possible. Are these managed switches, or just plug-n-go?
We've got about 10 plug and go, and 3 managed ones, but not much has been set from default on the managed ones. Some of the .pst files are above 1.5GB, but we've been getting some corrupted as small as .5GB.
All I can really suggest is to check the server that stores your user profiles to see if the event log has produced any SRV errors in the 2019 to 2022 range. Was the network 100Mb prior to the new switches and issues?
Can you tell if all the new corrupted boxes are on a particular switch?
All I can really suggest is to check the server that stores your user profiles to see if the event log has produced any SRV errors in the 2019 to 2022 range. Was the network 100Mb prior to the new switches and issues?
Yes the network was 100Mb prior to the new switches. I know we got the occasional corrupted mailbox, but not quite the food we are getting now.
Also, I was mistaken. We have 4 managed switces, and the are SMC brand. The rest are plug and go.
Can you tell if all the new corrupted boxes are on a particular switch?
Not for sure. We are still in the process of remapping our "organically grown through 3 IT managers" network. Most of the cabling seems good enough to run at 1Gb, and reterminating them most of the time will fix it when they don't run that fast. However probably on about 1/4 to 1/3rd of our computers actually have gigabit cards.
Yes the network was 100Mb prior to the new switches. I know we got the occasional corrupted mailbox, but not quite the food we are getting now.
Also, I was mistaken. We have 4 managed switces, and the are SMC brand. The rest are plug and go.
Not for sure. We are still in the process of remapping our "organically grown through 3 IT managers" network. Most of the cabling seems good enough to run at 1Gb, and reterminating them most of the time will fix it when they don't run that fast. However probably on about 1/4 to 1/3rd of our computers actually have gigabit cards.
Considering the roaming nature it may be impossible to tell, but any way to determine if the users having issues are only on the Gbit-enabled systems? Some switches have flow control negotiation issues in mixed networks.
Considering the roaming nature it may be impossible to tell, but any way to determine if the users having issues are only on the Gbit-enabled systems? Some switches have flow control negotiation issues in mixed networks.
Most but not all are on gigabit enabled computers. One of the ones I spent this morning trying to fix was definately not on a gigabit computer.
However, that does sound like a good reason to push the rest of the computers going to gigabit.
Ddruid_SMP
05-06-08, 09:19 PM
Check for Duplex and or Speed mis-matches. For the most part if the systems are stationary you are better off locking the port to that systems speed/duplex and turning auto-negotiation off. Some NIC's are fussy and will require to be locked as well.
dark_15
05-06-08, 11:56 PM
Can you span/mirror any of the ports on the managed switches? If so, it may be a good time to run a packet sniffer like wireshark (http://www.wireshark.org) to see if there any anomalies in communication between your server containing the profiles and host machines. Alternatively you could disable the PST files from being archived; however that means mail is going to stay local to the system that POP's off the mail. Are you using POP or IMAP to retrieve your mail?
Were using POP currently.
Boss decided to move the .pst files to a non-My Docs directory and they are being backed up by a script triggered robocopy at night. We'll see how that works.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.