September 29, 2009 at 9:30 am
We have a server running SQL2005 and is connected to a SAN. The RAID on the drives is optimal. Every once in a while we start getting this error:
The operating system returned error 1450(Insufficient system resources exist to complete the requested service.) to SQL Server during a write at offset 0x00000ff4f62000 in file with handle 0x0000000000001534. This is usually a temporary condition and the SQL Server will keep retrying the operation. If the condition persists then immediate action must be taken to correct it.
Followed by this.
ime out occurred while waiting for buffer latch -- type 2, bp 00000003AEFECB80, page 1:4756838, stat 0xc20009, database id: 8, allocation unit id: 72057594260553728, task 0x0000000003813828 : 0, waittime 4500, flags 0x1a, owning task 0x000000000384EB08. Continuing to wait.
This seems to completely freeze up the databases and the server. The only way we now how to fix it is by rebooting the server. Is this memory leak in the sql coding or something else that can be fixed?
September 29, 2009 at 9:54 am
SQL server doesn't have any memory leaks that I'm aware of here. Likely you are having IO issues. I've seen this when the firmware or drivers for the HBAs are not correct. I'd have this double checked to be sure that it's working correctly.
What do you mean by optimal? That's a nebulous term. If you are sharing drives behind the LUNs, it's possible there is IO contention.
September 29, 2009 at 10:16 am
I'm not talking about memory leaks in SQL I'm talking about memory leaks in the queries. I mean by optimal is the RAID says that there are no bad drives.
I have a Dell Blade Server connected to a Dell MD3000i SAN using Dells drivers. No other blade server is currently connected to this SAN.
Steve Jones - Editor (9/29/2009)
SQL server doesn't have any memory leaks that I'm aware of here. Likely you are having IO issues. I've seen this when the firmware or drivers for the HBAs are not correct. I'd have this double checked to be sure that it's working correctly.What do you mean by optimal? That's a nebulous term. If you are sharing drives behind the LUNs, it's possible there is IO contention.
September 29, 2009 at 10:24 am
The queries will not likely have memory leaks. You're using basic SQL that has been exhaustively tested. I guess it's possible, but I haven't seen it in years of doing this.
What I see often is that SAN admins don't have the right drivers, or the right versions on their HBAs. I've seen that over and over, with people refusing to check because they think these are the right drivers or firmware. That's why I suggest checking that first.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply