August 7, 2019 at 11:14 am
While restoring a log on a secondary replica to prepare a database for joining an availability group I suddenly received a flurry of 'There have been n misaligned log IOs which required falling back to synchronous IO' messages. In 20 years working with SQL Server this is a completely new message to me.
A Google search unearthed a few articles including the very recent https://www.sqlservercentral.com/blogs/error-misaligned-log-ios-which-required-falling-back-to-synchronous-io-and-slow-ag-synchronisation. However the fsutil output shows that sector sizes are exactly the same on both replicas, as might be expected on identically spec-ed AWS-hosted servers.
Any ideas?
August 7, 2019 at 2:45 pm
Couple things I found (since my search might be different than yours)
August 13, 2019 at 4:15 pm
We get these msgs on an AG where the sector sizes ("Bytes Per Physical Sector") are indeed different between datacenters. No easy solution, so we live with it. I have to say we have not really felt any impact, so we treat these msgs as informational.
Seems very curious that you get these msgs without that root cause...are you certain you have checked all of the data/log drives?
August 14, 2019 at 12:12 pm
Yes certain about the sector sizes being the same. As I said, these are disks on identically spec-ed AWS-hosted servers. It is indeed curious. I have not found any documentation about these messages that doesn't relate to unequal sector sizes. The messages do not seem to have had an impact and have not recurred. Hopefully it will stay that way.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply