Dmitrii Stepanov dstepanov-yadro
  • Joined on 2023-03-06
dstepanov-yadro pushed to master at TrueCloudLab/frostfs-node 2024-11-22 13:45:58 +00:00
1ed7ab75fb [#1517] cli: Print the reason of ape manager error
dstepanov-yadro merged pull request TrueCloudLab/frostfs-node#1517 2024-11-22 13:45:56 +00:00
Print the reason of ape manager error
dstepanov-yadro commented on pull request TrueCloudLab/frostfs-node#1476 2024-11-22 10:54:00 +00:00
engine: Optimize Inhume operation to improve speed with large object sets

I suggest to perform Inhume on different shards concurrently. Or I missing something?

dstepanov-yadro commented on pull request TrueCloudLab/frostfs-node#1476 2024-11-22 10:50:59 +00:00
engine: Optimize Inhume operation to improve speed with large object sets

I suggest to replace map[string][]int with map[string][]oid.Address. Anyway you do this in inhume method.

dstepanov-yadro commented on pull request TrueCloudLab/frostfs-node#1516 2024-11-21 11:07:39 +00:00
traverser: Check for placement vector out of range

I meant that the second slice is not updated.

dstepanov-yadro commented on pull request TrueCloudLab/frostfs-node#1516 2024-11-21 10:35:08 +00:00
traverser: Check for placement vector out of range

Looks like you have forgotten regularVector = append(regularVector, ns[i][rem[i]:]...) here

dstepanov-yadro pushed to feat/priority_queue at dstepanov-yadro/frostfs-node 2024-11-20 12:06:10 +00:00
17719116fa [#9999] quota: Add limiter
dstepanov-yadro commented on pull request TrueCloudLab/frostfs-observability#18 2024-11-20 09:13:53 +00:00
[#15] tracing: Add events for grpc stream calls

First of all, thanks for the PR. Аttachments are not opening. Can you attach images of what the events look like in jaeger? There is also an implementation proposal: add an event not only at…

dstepanov-yadro commented on pull request TrueCloudLab/frostfs-node#1476 2024-11-19 14:11:44 +00:00
engine: Optimize Inhume operation to improve speed with large object sets

To tell the truth, I don't like this approach. It turns out that one request with a large number of addresses to inhume can block other Inhume requests. I suggest using concurrent execution for each request.

dstepanov-yadro commented on issue TrueCloudLab/frostfs-node#1502 2024-11-18 14:03:43 +00:00
Be consistent in logging errors and use zap.Error

According to the issue on github, initially the problem was with errors from the pkg/errors, which stored information…