Re: CEPH S3 Listing Issue

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



I think this could be fixed by ListObjectV2.

On Mon, Mar 9, 2020 at 4:39 PM Daniel Gryniewicz <dgryniew@xxxxxxxxxx> wrote:
Certainly if you turn the RGW logging up to 20, you can see just about
everything that it does.  It should be possible to see the reason, but
of course there will be a lot of output in the logs.


Daniel

On 3/9/20 8:24 AM, rajiv chodisetti wrote:
> Any other pointers on how to debug this issue ? Will I be able to find
> these kind of issues through logs and if yes where can I find these
> logs , I have checked the storage gateway logs and I couldn't find any
> useful logs over there
>
> On Sat, Mar 7, 2020 at 9:35 AM rajiv chodisetti
> <rajivchodisetti54@xxxxxxxxx <mailto:rajivchodisetti54@xxxxxxxxx>> wrote:
>
>     HI Romit,
>
>     first of all thanks for your reply,
>
>     The Code that I have attached above fetches blobs in the batches
>     of 1000 because 1000 is what S3 client can fetch at a time.
>
>     * The total no of objects in bucket are around 3000 files and
>     total 2.8 GB in size
>     * To rule out the possibility of error in my code, I tried various
>     approaches in both java and as well as python (using boto) and by
>     increasing client timeout, retry limit, but it's the same
>     exception everywhere and I tried setting batch sizes to 10, 50,
>     100, 200,500 and 1000 but every other time listing fails after N
>     number of iterations, I have pasted the code snippets of various
>     approaches I have tried below
>     * I have increased the S3 gateway Pod count to rule out the
>     possibility of S3 gateway pod being bombaded
>     * I tried s3cmd recursive approach to get the file count of the
>     bucket and there also I see errors while listing
>
>     Please do let me know if there are any other ways I can check
>     what's going on at backend
>
>     Rajiv.Chodisetti@aakashonprem:~$ s3cmd ls -r s3://my-new-bucket/
>     --no-ssl --host=${AWS_HOST} --host-bucket=  s3://my-new-bucket >
>     listing.txt
>     WARNING: Retrying failed request:
>     /?marker=test-training-akash/du/latest/11340612-0.png.json
>     WARNING: 500 (UnknownError)
>     WARNING: Waiting 3 sec...
>     WARNING: Retrying failed request:
>     /?marker=test-training-akash/du/latest/11340612-0.png.json
>     WARNING: 500 (UnknownError)
>     WARNING: Waiting 6 sec...
>     ^CSee ya!
>
>
>     *    Approach 1:*
>                      for (Bucket bucket : s3.listBuckets()) {
>                     System.out.println(" - " + bucket.getName() + " "
>                             + "(owner = " + bucket.getOwner()
>                             + " "
>                             + "(creationDate = " +
>     bucket.getCreationDate());
>                     ObjectListing objectListing = s3.listObjects(new
>     ListObjectsRequest()
>                             .withBucketName(bucket.getName()));
>                     do {
>                         for (S3ObjectSummary objectSummary :
>     objectListing.getObjectSummaries()) {
>                             System.out.println(" --- " +
>     objectSummary.getKey() + " "
>                                     + "(size = " +
>     objectSummary.getSize() + ")" + " "
>                                     + "(eTag = " +
>     objectSummary.getETag() + ")");
>                         }
>                         objectListing =
>     s3.listNextBatchOfObjects(objectListing);
>                     } while (objectListing.isTruncated());
>                 }
>
>     *Approach 2:*
>         // where i is global variable to keep track of blobs listed so far
>         // Recursive code to list blobs in Batches of N
>         private void listDataset(Bucket bucket, AmazonS3 conn, String
>     nextMarker) {
>             try {
>                 ListObjectsRequest request = new ListObjectsRequest();
>                 request.setBucketName(bucket.getName());
>                 request.setMaxKeys(200);
>                 if (nextMarker != null && nextMarker.length() > 0) {
>                     request.setMarker(nextMarker);
>                 }
>                 ObjectListing result;
>                 result = conn.listObjects(request);
>                 for (S3ObjectSummary objectSummary :
>                         result.getObjectSummaries()) {
>                     System.out.println(" - " + objectSummary.getKey()
>     + "  " +
>                             "(size = " + objectSummary.getSize() +
>                             ")");
>                     i = i + 1;
>                 }
>                 System.out.println("Total Blobs Fetched So far: " + i);
>                 if (result.getNextMarker() != null) {
>                     System.out.println("Next Object Marker: " +
>     result.getNextMarker());
>                     nextMarker = result.getNextMarker();
>                     listDataset(bucket, conn, nextMarker);
>                 }
>             } catch (Exception e) {
>                 e.printStackTrace();
>             }
>         }
>
>     *  Approach 3:*
>
>         done = False
>         marker = None
>         while not done:
>             if marker:
>                 data = ""> >     MaxKeys=10, Marker=marker)
>             else:
>                 data = "" MaxKeys=10)
>
>             for key in data['Contents']:
>                 print(key)
>
>             if data['IsTruncated']:
>                 marker = data['Contents'][-1]['Key']
>             else:
>                 done = True
>
>     *Approach 4:*
>
>         paginator = s3.get_paginator('list_objects')
>         page_iterator = paginator.paginate(Bucket='my-new-bucket')
>
>         for page in page_iterator:
>              if "Contents" in page:
>                  for key in page["Contents"]:
>                      print(key)
>
>     On Sat, Mar 7, 2020 at 7:32 AM Romit Misra
>     <romit.misra@xxxxxxxxxxxx <mailto:romit.misra@xxxxxxxxxxxx>> wrote:
>
>         Some pointers:-
>
>         1.What is the approx object count in the bucket...?
>         2. Use range list,( iterate 1000 objects at a time, and store
>         the last marker to move to the next list, also referred to as
>         paginated listing)
>         3.Check your default timeouts, and optimize tune according to
>         environment
>         4.Monitor the system metrics when listing is happening, ( CPU
>         and memory)
>         5.Use a alternative client like s3cmd, or Boto rule out the
>         anamoly if in your code.
>
>
>         Thanks
>         Romit
>
>
>         On Sat, 7 Mar 2020, 02:34 rajiv chodisetti,
>         <rajivchodisetti54@xxxxxxxxx
>         <mailto:rajivchodisetti54@xxxxxxxxx>> wrote:
>
>             HI,
>
>             I have created an S3 bucket backed by CEPH and through
>             java S3 client and via S3 object gateway am listing all
>             the files under the bucket and always the listing is
>             failing some times after listing 1k+ blobs  or some times
>             after listing 2k+ blobs and am not able to figure out how
>             to debug this issue
>
>             This is the Exception am getting,
>
>             |com.amazonaws.services.s3.model.AmazonS3Exception: null
>             (Service: Amazon S3; Status Code: 500; Error Code:
>             UnknownError; Request ID:
>             tx00000000000000000e7df-005e626049-1146-rook-ceph-store;
>             S3 Extended Request ID:
>             1146-rook-ceph-store-rook-ceph-store), S3 Extended Request
>             ID: 1146-rook-ceph-store-rook-ceph-store at
>             com.amazonaws.http.AmazonHttpClient$RequestExecutor.handleErrorResponse(AmazonHttpClient.java:1799)|
>
>
>             I tried with Boto as well and it's the same error over there,
>
>             I have checked the s3 gateway pod logs but I couldn't find
>             any relevant logs over there, so, please let me know how
>             do I debug this issue or possible reasons for the same
>
>             I have attached the java code am using for reference.
>
>             Thanks & Regards,
>             Rajiv
>             UIPath
>             _______________________________________________
>             Dev mailing list -- dev@xxxxxxx <mailto:dev@xxxxxxx>
>             To unsubscribe send an email to dev-leave@xxxxxxx
>             <mailto:dev-leave@xxxxxxx>
>
>
>>         /-----------------------------------------------------------------------------------------/
>>
>>         /This email and any files transmitted with it are
>>         confidential and intended solely for the use of the
>>         individual or entity to whom they are addressed. If you have
>>         received this email in error, please notify the system
>>         manager. This message contains confidential information and
>>         is intended only for the individual named. If you are not the
>>         named addressee, you should not disseminate, distribute or
>>         copy this email. Please notify the sender immediately by
>>         email if you have received this email by mistake and delete
>>         this email from your system. If you are not the intended
>>         recipient, you are notified that disclosing, copying,
>>         distributing or taking any action in reliance on the contents
>>         of this information is strictly prohibited./
>>
>>         /Any views or opinions presented in this email are solely
>>         those of the author and do not necessarily represent those of
>>         the organization. Any information on shares, debentures or
>>         similar instruments, recommended product pricing, valuations
>>         and the like are for information purposes only. It is not
>>         meant to be an instruction or recommendation, as the case may
>>         be, to buy or to sell securities, products, services nor an
>>         offer to buy or sell securities, products or services unless
>>         specifically stated to be so on behalf of the Flipkart group.
>>         Employees of the Flipkart group of companies are expressly
>>         required not to make defamatory statements and not to
>>         infringe or authorise any infringement of copyright or any
>>         other legal right by email communications. Any such
>>         communication is contrary to organizational policy and
>>         outside the scope of the employment of the individual
>>         concerned. The organization will not accept any liability in
>>         respect of such communication, and the employee responsible
>>         will be personally liable for any damages or other liability
>>         arising./
>>
>>         /Our organization accepts no liability for the content of
>>         this email, or for the consequences of any actions taken on
>>         the basis of the information /provided,/ unless that
>>         information is subsequently confirmed in writing. If you are
>>         not the intended recipient, you are notified that disclosing,
>>         copying, distributing or taking any action in reliance on the
>>         contents of this information is strictly prohibited./
>>
>>         /-----------------------------------------------------------------------------------------/
>>
>
> _______________________________________________
> Dev mailing list -- dev@xxxxxxx
> To unsubscribe send an email to dev-leave@xxxxxxx
_______________________________________________
Dev mailing list -- dev@xxxxxxx
To unsubscribe send an email to dev-leave@xxxxxxx
_______________________________________________
Dev mailing list -- dev@xxxxxxx
To unsubscribe send an email to dev-leave@xxxxxxx

[Index of Archives]     [CEPH Users]     [Ceph Devel]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux