What is meant by "properly handles writing cached data to persistent disks when needed" #39740
Comments
|
Thanks for the feedback! We are currently investigating and will update you shortly. |
|
Yes you can enable it for SQL server. SQL is able to understand and properly write the cached data to persistent storage. This is built into the design of SQL. However if you are running a non SQL application, that ability may not be built into your application and in turn could result in data loss if the VM crashes. So this statement is simply saying, you can enable read/write caching on data disks associated with a SQL application but for non SQL apps you might hit issues unless your app is configured properly. |
|
Hi, thanks for this clarification. Still, it’s unclear what an application
needs to do, e.g. are fsyncs or windows equivalant respected?
Ideally this page would describe the syscalls you need to do to flush the
cache to stable storage.
…On Tue, 1 Oct 2019 at 00:37, Micah ***@***.***> wrote:
Yes you can enable it for SQL server. SQL is able to understand and
properly write the cached data to persistent storage. This is built into
the design of SQL. However if you are running a non SQL application, that
ability may not be built into your application and in turn could result in
data loss if the VM crashes.
So this statement is simply saying, you can enable read/write caching on
data disks associated with a SQL application but for non SQL apps you might
hit issues unless your app is configured properly.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#39740?email_source=notifications&email_token=AA35X2I246E64DQ645G52XLQMJ5UDA5CNFSM4I3R35W2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD77KAUQ#issuecomment-536780882>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AA35X2K7PULMJA6DODFNTGLQMJ5UDANCNFSM4I3R35WQ>
.
|
|
The doc is focused around SQL so I would assume we don't want to add extra information around this. However I will assign to the content authors to review. |
mimckitt
added
assigned-to-author
doc-enhancement
and removed
cxp
in-progress
product-question
labels
|
How do you figure that? The doc is about VM disks in general, I think SQL
is only mentioned as an example.
It’s important for other databases as well to know which cache settings you
can use when running them.
…On Tue, 1 Oct 2019 at 19:35, Micah ***@***.***> wrote:
The doc is focused around SQL so I would assume we don't want to add extra
information around this. However I will assign to the content authors to
review.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#39740?email_source=notifications&email_token=AA35X2NBQRYSE2NGVT6SAUDQMOC55A5CNFSM4I3R35W2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEACC77A#issuecomment-537145340>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AA35X2KF5NDNTMEI3GHC2JTQMOC55ANCNFSM4I3R35WQ>
.
|
|
|
|
But I wouldn't disagree. Likely should include more info or avoid just using SQL as an example throughout the whole thing |
|
Yes, agreed. Just as a side note, currently for me the fact that this info
is lacking prevents me from selecting Azure VMs as a viable storage option
for databases since the implications of the cache storage settings with
regard to durability and what an application needs to do to make it durable
are - as far as I can tell - undocumented.
…On Tue, 1 Oct 2019 at 19:44, Micah ***@***.***> wrote:
But I wouldn't disagree. Likely should include more info or avoid just
using SQL as an example throughout the whole thing
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#39740?email_source=notifications&email_token=AA35X2KMLHMI5UGIS3KLEUDQMOD7VA5CNFSM4I3R35W2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEACD5XQ#issuecomment-537149150>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AA35X2JIE2GW6PARBZ7W373QMOD7VANCNFSM4I3R35WQ>
.
|
|
#assign: @MicahMcKittrick-MSFT |
|
We are currently cleaning up old issues and closing out items that are greater than 90 days old. If an issue is still present, please open a new feedback item on the document so we can prioritize correctly. #please-close |
|
In case anyone is looking at this, it seems you need to use FILE_FLAG_WRITE_THROUGH flag when doing writes: https://azure.microsoft.com/en-us/blog/data-series-exploring-windows-azure-drives-disks-and-images/ The article states that the write cache is only in memory. |
The docs say: "Configure host-cache as ReadWrite only if your application properly handles writing cached data to persistent disks when needed.". But it's unclear what properly means in this context. The example for SQL Server: "For example, SQL Server handles writing cached data to the persistent storage disks on its own. Using ReadWrite cache with an application that does not handle persisting the required data can lead to data loss, if the VM crashes.", doesn't make things much clearer: can we enable it for SQL Server or not?
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
The text was updated successfully, but these errors were encountered: