[Disk Manager] sync test/acceptance

cloud/disk_manager/test/acceptance/README.md

@@ -0,0 +1,53 @@
+# Acceptance test
+
+## Description
+1. Runs main test:
+    1. Creates snapshot from the disk (via `ycp`), then creates incremental snapshot and disk from this snapshot (via `ycp`).
+    2. Creates pooled image from snapshot (via `ycp`), then creates disk from this image (via `ycp`).
+    3. Creates image from disk, created on step 1.1 .(via `ycp`), then creates image from this image (via `ycp`).
+2. Runs cancel test:
+    1. Asynchronously creates snapshot for every disk (via `ycp`).
+    2. Tries to delete snapshot while it is in CREATING status (via `ycp`).
+    3. Checks that snapshot was successfully deleted (via `ycp`).
+
+## Usage
+Install `ycp` and reload shell:
+```(bash)
+$ curl https://s3.mds.yandex.net/mcdev/ycp/install.sh | bash
+$ cat > ~/.bashrc <<EOF
+PATH="$HOME/ycp/bin/:$PATH"
+EOF
+```
+
+Install `ycp` configuration for tests:
+```(bash)
+$ mkdir -p ~/.config/ycp/
+$ mv ~/.config/ycp/config.yaml ~/.config/ycp/config.yaml.bak
+$ cp ../../../packages/yc-nbs-ci-tools/ycp-config.yaml ~/.config/ycp/config.yaml
+```
+
+Patch `/etc/hosts` to be able to use `ycp` with `hw-nbs-stable-lab`:
+```(bash)
+$ export HW_NBS_STABLE_LAB_SEED_IP=$(host $(pssh list C@cloud_hw-nbs-stable-lab_seed) | awk '{print $5}')
+$ echo -e "\n# ycp hack for hw-nbs-stable-lab\n$HW_NBS_STABLE_LAB_SEED_IP local-lb.cloud-lab.yandex.net" | sudo tee -a /etc/hosts
+```
+
+Create `/etc/ssl/certs/hw-nbs-stable-lab.pem` with the content https://paste.yandex-team.ru/3966169
+
+Save private ssh key from https://yav.yandex-team.ru/secret/sec-01ehpt7c9ez5g4j9nx4g4yegj3/explore/version/ver-01ehpt7c9ng1t28517aqjpavd5 to ~/.ssh/overlay, then `ssh-add ~/.ssh/overlay`
+
+Build `acceptance-test`:
+```(bash)
+$ ya make
+```
+
+Example:
+```(bash)
+$ ./acceptance --profile hw-nbs-stable-lab --folder-id yc.dogfood.serviceFolder --src-disk-ids <any_disk_id>,<another_disk_id> --verbose
+```
+
+## Leaked resources
+There is a possibility when the test can leave leaked resources inside the cloud. So if it is happened:
+1. Find all snapshots inside the cluster with the name pattern `acceptance-test-snapshot-<suffix>-<timestamp>` and delete them manually.
+2. Find all images inside the cluster with the name pattern `acceptance-test-image-<suffix>-<timestamp>` and delete them manually.
+3. Find all disks inside the cluster with the name pattern `acceptance-test-disk-<suffix>-<timestamp>` and delete them manually.

cloud/disk_manager/test/acceptance/cmp/main.cpp

@@ -0,0 +1,98 @@
+#include <library/cpp/getopt/last_getopt.h>
+#include <library/cpp/logger/log.h>
+
+#include <util/datetime/base.h>
+#include <util/string/builder.h>
+#include <util/system/file.h>
+
+////////////////////////////////////////////////////////////////////////////////
+
+struct TParams
+{
+    bool Verbose = false;
+    size_t Bytes = 0;
+    size_t IgnoreInitial = 0;
+    size_t ChunkSize = 4 * 1024 * 1024;
+    int MaxRetriesAfterFailure = 5;
+    TString File1;
+    TString File2;
+};
+
+////////////////////////////////////////////////////////////////////////////////
+
+int Run(const TParams& params)
+{
+    TLog log(CreateLogBackend("cout"));
+    log.SetFormatter([] (ELogPriority, TStringBuf msg) {
+        return TStringBuilder() << "[" << TInstant::Now().FormatLocalTime("%Y-%m-%d %H:%M:%S") << "] " << msg;
+    });
+
+    auto f1 = TFile(params.File1, RdOnly | DirectAligned);
+    auto f2 = TFile(params.File2, RdOnly | DirectAligned);
+
+    f1.Seek(params.IgnoreInitial, sSet);
+    f2.Seek(params.IgnoreInitial, sSet);
+
+    TVector<ui8> buf1(params.ChunkSize);
+    TVector<ui8> buf2(params.ChunkSize);
+
+    int retriesAfterFailure = 0;
+    size_t offset = 0;
+    while (offset < params.Bytes) {
+        f1.Load(buf1.data(), buf1.size());
+        f2.Load(buf2.data(), buf2.size());
+
+        bool failed = false;
+        for (size_t i = 0; i < buf1.size(); i++) {
+            if (buf1[i] != buf2[i]) {
+                log << offset + i << " " << (ui32)buf1[i] << " " << (ui32)buf2[i] << Endl;
+                failed = true;
+            }
+        }
+
+        if (failed) {
+            if (retriesAfterFailure < params.MaxRetriesAfterFailure) {
+                retriesAfterFailure++;
+                log << "files mismatch, retry " << retriesAfterFailure << Endl;
+
+                f1.Seek(-params.ChunkSize, sCur);
+                f2.Seek(-params.ChunkSize, sCur);
+                continue;
+            }
+
+            log << "max number of retries after failure reached, abort" << Endl;
+            return 1;
+        }
+
+        if (retriesAfterFailure > 0) {
+            log << "retry after failure succeeded at offset " << offset << Endl;
+            return 1;
+        }
+
+        offset += params.ChunkSize;
+    }
+
+    return 0;
+}
+
+////////////////////////////////////////////////////////////////////////////////
+
+int main(int argc, char** argv)
+{
+    TParams params;
+
+    NLastGetopt::TOpts opts;
+    opts.AddLongOption("verbose").NoArgument().StoreResult(&params.Verbose);
+    opts.AddLongOption("bytes").StoreResult(&params.Bytes);
+    opts.AddLongOption("ignore-initial").StoreResult(&params.IgnoreInitial);
+    opts.AddLongOption("chunk-size").StoreResult(&params.ChunkSize);
+    opts.AddLongOption("max-retries-after-failure").StoreResult(
+        &params.MaxRetriesAfterFailure);
+    opts.AddFreeArgBinding("file1", params.File1);
+    opts.AddFreeArgBinding("file2", params.File2);
+    opts.SetFreeArgsMax(2);
+
+    NLastGetopt::TOptsParseResult r(&opts, argc, argv);
+
+    return Run(params);
+}

cloud/disk_manager/test/acceptance/cmp/ya.make

@@ -0,0 +1,14 @@
+OWNER(g:cloud-nbs)
+
+PROGRAM(acceptance-cmp)
+
+SRCS(
+    main.cpp
+)
+
+PEERDIR(
+    library/cpp/getopt/small
+    library/cpp/logger
+)
+
+END()