Welcome to our community

Any help on this as I am scratching my head like mad,
installed an MSA1510i with two targets, first SQL target is araid 1+0 stripe 16kb, runs smooth and no problems, linked to DL360 G4 on its own dual switched network, other target is a file server on a raid 5 128kb stripe, again dl360 going through same switches, full redundancy and servers have two gig ethernet(TOE compliant cards) to the MSA and one to LAN

file replciation using storage mirroring on both targets to a storage server on the LAN.
MPIO used for round robin to MSA on both servers.

now SQL runs fine and dandy, no issues however the file target seems to halt every now and then and hang for 20 seconds or more, then it speeds up again.

there can be as little as one user and still does the same.


any help?

thanks in advance, balding in patches now!

Hello Sleazy,

Welcome to the forum
- do you have a bit more info - such as the type of file server - is it a dedicated HP storage server ? do you get any problems with the NIC that connects to the LAN ? are you using physically seperate networks, or VLAN's ?

Have you tried to "move" the problem by maybe forcing traffic down one NIC, then the other - or would that cause too many issues on site ?

FT

Thanks for the reply,
there is two HP DL360's multihomed on the front end, both have dual nics into the MSA through a two netgear switch's, it seems this has now progressed and one of the DL360's is having a similar problem through the LAN, it is running W2k server, the other server is W2003, I am in the process of downloading the latest Maintainance pack and looking to see what updates are out there as I get problems running Chdsk, w2003 box is fine, also does not reconnect to the MSA on boot correctly however the w2003 box does, both using same version of MPIO & MS initiator.


bit stumped really!


thanks for the help

Hello Sleazy,

From what you say - it sounds like the common denominator is the W2K box - network hangs/pauses are commonly caused by things like spanning tree, or routing recalculations. Sounds like some more investigation is needed - if the Netgear switches are managed - it might be possible to capture some traces, or console output at the time of the problem. The same goes for the W2K box - you can use a utility like Ethereal to capture network traffic. An intermittant NIC problem could be causing paths to change.

Basically, its down to basic troubleshooting analysis to try and identify the failure. So - Whats changed/different, Where is the problem located, What is the extent of the problem, and when did/does it occur ?.

Let us know how this develops.

FT

Thanks ,
I have started to perf mon on the network and am running ethereal, chdsk seems to find errors and clear them and it speeds up.

I think the only option to clear the lot in one go is to upgrade to 2003 server. the shares dont seem to reconnect on reboot and other things manifasting now which are a bit too vague.

if it was both servers I would rethink my config however am convinced now its 2k with the issues.

thanks for your help