How to handle exception with blob_clent_wrapper

[Gaurav-Karu]

Blob_clent_wrapper return void ....and If I go with blob_clent the volume of download I cannot achieve concurrency....like parallel upload.

Please suggest we need to integrate cpplite as azure-iot-sdk do not have a support for proxy.

Please share some sample code as sample2.cpp doesn't seem to be enough

[Jinming-Hu]

@Gaurav-Karu hi, cpplite is going to be deprecated in favor of Track2 storage SDK.

[Gaurav-Karu]

@Jinming-Hu : Yes I am aware of that as it has been conveyed to us by Swadi.Shraddha@microsoft.com , But Does the Track2 storage is in production ?

Note: we have already given the dates to client ....so what I can do and we cannot experiments until and unless the new SDK is under production ?

we are switching from azure-iot-sdk-c to azure-cpplite because cpplite has a support for proxy

[Jinming-Hu]

blob_client_wrapper doesn't throw exception, you'll need to check errno for error information.

[Jinming-Hu]

and blob_client does have concurrent upload function

azure-storage-cpplite/include/blob/blob_client.h

Line 154 in be490ed

AZURE_STORAGE_API std::future<storage_outcome<void>> upload_block_blob_from_buffer(const std::string &container, const std::string &blob, const char* buffer, const std::vector<std::pair<std::string, std::string>> &metadata, uint64_t bufferlen, int parallelism = 1);

[Jinming-Hu]

@Jinming-Hu : Yes I am aware of that as it has been conveyed to us by Swadi.Shraddha@microsoft.com , But Does the Track2 storage is in production ?

Not yet, but it's the latest beta8 is considered RC release. So there won't be too many breaking changes between this version and production release (GA).

[Gaurav-Karu]

@Jinming-Hu ...Is there a way I can recompile cpplite and return errno in another wrapper class which I can use in my code to check the errno....because using assert is dangerous in the production code.

Can u please suggest ?

[Jinming-Hu]

You don't have to use asserts.

Just check errno like this

client.create_container(your_container_name);
int ret = errno;
if (ret == 0) {
    // succeeded
} else if (ret == container_already_exists) {
    // container already exists
}

there's a list of error numbers here

or if you want to wrap code above into a function

int create_container_wrapper(const std::string& container_name) {
    client.create_container(container_name);
    return errno;
}

[Gaurav-Karu]

Hi Thanks for the update...I have one more query if the files are larger size (Let say 800MB) it will tgake some time to upload in blob storage....Now if the user suspend the upload and resume back. (Do such handling is there ?)

[Gaurav-Karu]

why blob_client_wrapper obj is not made accessible directly ? And do we need to use only one object blob_client_wrapper to handle upload, Download and delete or we can create different blob_client_wrapper object for handling ?
How to make use of it ?

[Jinming-Hu]

I have one more query if the files are larger size (Let say 800MB) it will tgake some time to upload in blob storage....Now if the user suspend the upload and resume back. (Do such handling is there ?)

Nope, there's no way to suspend or resume for now. But there is a workaround. We have Put Block and Put Block List API. You can split that 800MB data into small chunks, and upload one or a few chunks at a time.

why blob_client_wrapper obj is not made accessible directly ? And do we need to use only one object blob_client_wrapper to handle upload, Download and delete or we can create different blob_client_wrapper object for handling ?
How to make use of it ?

blob_client_wrapper is not for public use, it was designed for a specific customer. So the API wasn't well-designed. That's why I 've been suggesting you take a look at our new Track2 SDK. There's already an RC-release. The API interface is quite stable.

You can use single client_wrapper for multiple operations, even perform these operations from multiple threads.

[Gaurav-Karu]

How to get the size of committed and uncommitted block if I am using blob_client_wrapper function upload_file_to_blob

[Jinming-Hu]

@Gaurav-Karu take a look at this API

[Gaurav-Karu]

what is the alternative for this api in cpplite ?
///

/// Returns an enumerable collection of the blob's blocks, using the specified block list filter.
///

/// One of the enumeration values that indicates whether to return
/// committed blocks, uncommitted blocks, or both.
/// An object that represents the access condition for the operation.
/// An object that specifies additional options for the request.
/// An object that represents the context for the current operation.
/// An enumerable collection of objects implementing .
std::vector<block_list_item> download_block_list(block_listing_filter listing_filter, const access_condition& condition, const blob_request_options& options, operation_context context) const
{

[Jinming-Hu]

get_block_list

[Gaurav-Karu]

    /// <summary>
    /// Uploads a list of blocks to a new or existing blob. 
    /// </summary>
    /// <param name="block_list">An enumerable collection of block IDs, as Base64-encoded strings.</param>
    /// <param name="condition">An <see cref="azure::storage::access_condition" /> object that represents the access condition for the operation.</param>
    /// <param name="options">An <see cref="azure::storage::blob_request_options" /> object that specifies additional options for the request.</param>
    /// <param name="context">An <see cref="azure::storage::operation_context" /> object that represents the context for the current operation.</param>
    void upload_block_list(const std::vector<block_list_item>& block_list, const access_condition& condition, const blob_request_options& options, operation_context context)
    {
        **upload_block_list_async**(block_list, condition, options, context).wait();
    }

Do I need to call upload again or any specific api in cpplite who does similar ?