参考资料:Anthropic Context Rot 博客、Chroma Context Rot 研究、Stanford Lost in the Middle 论文、NVIDIA RULER 基准测试、ICML 2025 Unable to Forget 论文、NoisyBench、Medium “Your Context Window Is Lying to You”、arXiv 多项长上下文退化研究。
Jul 7 16:48:16 kernel: Out of memory: Kill process 1983 (httpd) score 27
Jul 7 16:48:23 kernel: Out of memory: Kill process 1985 (httpd) score 27
Jul 7 16:48:28 kernel: Out of memory: Kill process 1987 (httpd) score 27
...(每 3 秒被杀一个,连续 10+ 个)
Hinton 2015, arXiv:1503.02531。Claude Opus 4.8强Agent/Coding(Terminus-2 84)弱HLE数学(41.4),GPT-5.5强综合推理(HLE 45)弱Agent/Coding(SWE Pro 54.2),单蒸任一个都出不了GLM-5.2的HLE 40.5 + SWE Pro 62.1 + Terminus-2 81组合
Kaplan J, McCandlish S, Henighan T, et al. Scaling laws for neural language models. arXiv:2001.08361, 2020.
[2]
Hoffmann J, Borgeaud S, Mensch A, et al. Training compute-optimal large language models. arXiv:2203.15556, DeepMind, 2022.
[3]
Hinton G, Vinyals O, Dean J. Distilling the knowledge in a neural network. arXiv:1503.02531, 2015.
[4]
Shumailov I, Shumaylov Z, Zhao Y, et al. AI models collapse when trained on recursively generated data. Nature, 2024.
[5]
Gunasekar S, Tsipras D, et al. Textbooks are all you need. Microsoft Research, 2023.
[6]
Zhou C, Liu P, Xu P, et al. LIMA: less is more for alignment. Meta, 2023.
[7]
DeepSeek-AI. DeepSeekMath: pushing the limits of mathematical reasoning in open language models. arXiv:2402.03300, 2024.
[8]
DeepSeek-AI. DeepSeek-R1 technical report. 2024.
[9]
Villalobos P, Sevilla J, Heim L, et al. Will we run out of data? An analysis of the limits of scaling datasets in machine learning. arXiv:2211.04325, Epoch AI, ICML 2024.
[10]
Penedo G, et al. FineWeb: decanting the web for high-quality LLM pretraining data. Hugging Face, 2024.
[11]
Li Y, et al. DCLM: data-centric LLM benchmarking via filtered common crawl. 2024.
二、行业研报与机构
序号
文献信息
[12]
Patel D, Wong G. GPT-4 architecture, infrastructure, training dataset, costs, vision, MoE. SemiAnalysis, 2023-07-10.
[13]
Epoch AI. Will we run out of data? Limits of LLM scaling based on human-generated text. 2024 update.
[14]
Graphite. State of AI content 2025: 52% of new English articles AI-generated. 2025.
[15]
Ahrefs. AI content in the web: 74.2% of new pages contain AI content. 2025.
[16]
Imperial College London, Stanford, Internet Archive. Estimating AI-generated content on the public web: ~35% of global new content with AI traces. 2025.
三、公开披露
序号
文献信息
[17]
北京智谱华章科技股份有限公司.首次公开发行股票招股说明书(申报稿).香港交易所,2024.
[18]
OpenAI. How OpenAI’s models influenced DeepSeek’s performance and safety standards. OpenAI Official Blog, 2024-02.
[19]
Amodei D. Interview on Chinese models distilling overseas API data. Wall Street Journal, 2024-05.
[20]
OpenAI. Terms of service, Sec 2.2: restrictions on training competing models. 2023-11 update.
[21]
Kang D. Human data is (probably) more expensive than compute for training frontier LLMs. Substack, 2025-08.
ERROR - Unexpected exception occurred invoking async method:
DictionaryVectorRepositoryImpl.remove()
java.lang.IllegalArgumentException: collectionName is blank
at CollectionRequest$Builder.build(CollectionRequest.java:87)
at VectorDb.exists(VectorDb.java:125)
at VectorDb.deleteByCondition(VectorDb.java:256)
at DictionaryVectorRepositoryImpl.remove(DictionaryVectorRepositoryImpl.java:189)
核心错误:collectionName is blank — 向量集合名称为空。
第三步:检查数据库配置
错误明确指向集合名称为空,直接查数据库确认:
SELECT dictionary_id, dictionary_code, collection_name, sync_type
FROM hzero_aip.haip_aigc_dictionary;
UPDATE hzero_aip.haip_aigc_dictionary
SET collection_name = 'haip_dictionary_value_0'
WHERE dictionary_id IN (2, 3);
完整修复步骤
1. 执行SQL补全字段
docker exec mysql-hzero mysql -uhzero -phzero -e "
UPDATE hzero_aip.haip_aigc_dictionary
SET collection_name = 'haip_dictionary_value_0'
WHERE dictionary_id IN (2, 3);"
ERROR --- Failed on node AIGC 生成低代码业务对象和页面/LOWCODE.BUSINESS_OBJECT_PAGE/INTERNAL_INTERFACE_...
io.choerodon.core.exception.CommonException: 程序出现错误,请联系管理员
at ResponseUtils.getResponse(ResponseUtils.java:133)
at AbstractInnerRequestNodeHandler.execute(AbstractInnerRequestNodeHandler.java:199)
MissingServletRequestParameterException: Required request parameter 'domainCode' for method parameter type String is not present
Request: {URI=/v1/0/business-objects/domain/simple-generator}
A “Hello, world” program is usually a simple computer program that displays on the screen a message similar to “Hello, world”. A small piece of code in most general-purpose programming languages, this program is used to illustrate a language’s…
type: Computer program
source: Wikipedia
测试正常
Step4 看Openclaw技能列表状态:
openclaw skills list
✓ ready │ 📦 anysearch │ Real-time search engine supporting web search, vertical domain search, parallel batch search, and URL content extraction. │ openclaw-managed │
状态正常
from pptx import Presentation
import os
# 案例素材
case\_data = {
'客户名称': '{客户名称}',
'标题': '标杆案例-{客户名称}{场景名称}智能体',
'客户行业地位': '{行业地位,≤40字符}',
'收益亮点': '{收益亮点,≤40字符,用↑↓}',
'业务域': '{业务域}',
'痛点': \[
'{痛点一,≤25字符}',
'{痛点二,≤25字符}',
'{痛点三,≤25字符}',
'{痛点四,≤25字符}'
],
'收益': \[
'{收益一,≤25字符,含数字}',
'{收益二,≤25字符}',
'{收益三,≤25字符}',
'{收益四,≤25字符}'
],
'解决方案': \[
('{方案1标题}', '{方案1描述,≤30字符}'),
('{方案2标题}', '{方案2描述,≤30字符}'),
('{方案3标题}', '{方案3描述,≤30字符}'),
('{方案4标题}', '{方案4描述,≤30字符}'),
('{方案5标题}', '{方案5描述,≤30字符}')
]
}
# 加载模板
template\_path = 'AIGC平台材料/汉得信息H-AI平台应用案例集-汇总-模板-Sharp.pptx'
output\_path = 'AIGC平台材料/{客户名称}-{场景名称}案例.pptx'
prs = Presentation(template\_path)
slide = prs.slides\[0]
# 替换段落文本函数
def set\_paragraph\_text(paragraph, new\_text):
runs = paragraph.runs
if runs:
runs\[0].text = new\_text
for run in runs\[1:]:
run.text = ''
# 收集所有段落
all\_paragraphs = \[]
for shape in slide.shapes:
if shape.has\_text\_frame:
for para in shape.text\_frame.paragraphs:
all\_paragraphs.append(para)
# 段落索引替换映射
para\_replacements = {
0: case\_data\['标题'],
1: f"一句话客户行业和地位:{case\_data\['客户行业地位']}",
2: f"客户应用后收益亮点:{case\_data\['收益亮点']}",
4: f"客户在{case\_data\['业务域']}业务域有以下痛点:",
7: f"痛点一:{case\_data\['痛点']\[0]}",
9: f"痛点二:{case\_data\['痛点']\[1]}",
11: f"痛点三:{case\_data\['痛点']\[2]}",
13: f"痛点四:{case\_data\['痛点']\[3]}",
16: "应用后有如下收益:",
19: f"收益一:{case\_data\['收益']\[0]}",
21: f"收益二:{case\_data\['收益']\[1]}",
23: f"收益三:{case\_data\['收益']\[2]}",
25: f"收益四:{case\_data\['收益']\[3]}",
37: case\_data\['客户名称'],
38: "",
31: "",
}
# 执行替换
for para\_idx, new\_text in para\_replacements.items():
if para\_idx < len(all\_paragraphs):
set\_paragraph\_text(all\_paragraphs\[para\_idx], new\_text)
# 替换解决方案区域
solution\_idx = 0
for para in all\_paragraphs:
full\_text = ''.join(\[run.text for run in para.runs])
if '示意图' in full\_text and solution\_idx < 5:
set\_paragraph\_text(para, case\_data\['解决方案']\[solution\_idx]\[0])
solution\_idx += 1
elif '核心步骤' in full\_text and solution\_idx <= 5:
idx = solution\_idx - 1
if idx < 5:
set\_paragraph\_text(para, case\_data\['解决方案']\[idx]\[1])
# 保存
prs.save(output\_path)
print(f"✅ PPT生成成功: {output\_path}")
—
验证检查清单
生成后必须验证:
import zipfile
with zipfile.ZipFile(output\_path, 'r') as z:
files = z.namelist()
# 1. 母版检查
assert any('slideMaster' in f for f in files), "母版丢失"
# 2. 主题检查
assert any('theme' in f for f in files), "主题丢失"
# 3. 文件大小检查
file\_size = os.path.getsize(output\_path)
assert file\_size > 100000, f"文件过小: {file\_size}"
print("✅ 所有验证通过")
$ obsidian-cli --help
Interact with Obsidian vaults from the terminal
Available Commands:
add-vault Register a vault directory
create Creates note in vault
daily Creates or opens daily note
delete Delete note in vault
frontmatter View or modify note frontmatter
list List files and folders
list-vaults Lists all registered vaults
move Move or rename note (auto-update links)
open Opens note by name
print Print contents of note
remove-vault Unregister a vault
search Fuzzy search notes
search-content Search note content
set-default-vault Sets default vault
前置需求(SKILL中已经说明,Agent读取SKILL应该会自动安装):
ffmpeg – For audio extraction from video files
dashscope (Python SDK) – For SDK transcription method
requests – For async API method (这个要在openclaw的配置文件openclaw.json 的模型配置中有dashscope 或者 aliyun名称的模型供应商名称)
实际测试效果:
指令:总结视频内容 : videos/单据复刻片段1.mp4
Agent输出总结内容:
======================
用户提供了对话的元数据信息,现在可以生成完整的视频总结。我需要整合:
视频基本信息
视觉分析结果(关键帧)
音频转录内容
综合总结
这是一个企业级演示视频,展示了通过 AI 多模态模型(Kimi K2.5)从其他系统截图自动生成 HZERO 低代码平台业务对象和页面的完整流程。
让我生成一个完整的视频总结。## 🎬 视频内容总结:HZERO 单据页面智能复刻演示
📋 基本信息
属性
值
视频时长
633.5 秒 (10分33秒)
分辨率
752×1036 (竖屏)
关键帧数
20 帧
转录字符
2073 字符
处理速度
16.9x 实时 (FunASR)
🎯 视频核心内容
主题: 展示 Kimi K2.5 多模态模型 + HZERO AI 平台实现单据页面智能复刻的完整流程
Gateway: bind=loopback (127.0.0.1), port=18789 (service args) Probe target: ws://127.0.0.1:18789 Dashboard: http://127.0.0.1:18789/ Probe note: Loopback-only gateway; only local clients can connect.
Runtime: running (pid 389, state active, sub running, last exit 0, reason 0) RPC probe: ok
Listening: 127.0.0.1:18789 Troubles: run openclaw status Troubleshooting: https://docs.openclaw.ai/troubleshooting root@desktop-jacksen:# ^C root@desktop-jacksen:#
Openclaw控制台也能正常访问,而且看飞书渠道的 H0-项目开发经理 机器人状态也正常。
但就是没响应。
看Gateway的问题:
openclaw gateway probe
root@desktop-jacksen:# openclaw gateway probe
�� OpenClaw 2026.3.13 (61d171a) — Self-hosted, self-updating, self-aware (just kidding… unless?). ◇ Gateway Status Reachable: no Probe budget: 3000ms
Discovery (this machine) Found 0 gateways via Bonjour (local.) Tip: if the gateway is remote, mDNS won’t cross networks; use Wide-Area Bonjour (split DNS) or SSH tunnels.
Targets Local loopback ws://127.0.0.1:18789 Connect: failed – timeout
root@desktop-jacksen:#
结果显示Gateway 不可达,超时了,不健康; 但过了较长一段时间看飞书机器状态,又有回应了:
看起来,在问他问题的时候,他正在做自己以前的任务,没空搭理我。
运行 open claw诊断医生:
openclaw doctor
oot@desktop-jacksen:# openclaw doctor
�� OpenClaw 2026.3.13 (61d171a) — I autocomplete your thoughts—just slower and with more API calls.
�� OpenClaw 2026.3.13 (61d171a) — Runs on a Raspberry Pi. Dreams of a rack in Iceland. gateway connect failed: Error: gateway closed (1000): Gateway not reachable. Is it running and accessible? Gateway target: ws://127.0.0.1:18789 Source: local loopback Config: /root/.openclaw/openclaw.json Bind: loopback Hint: run `openclaw doctor` root@desktop-jacksen:# openclaw gateway status
�� OpenClaw 2026.3.13 (61d171a) — Your config is valid, your assumptions are not.
Runtime: running (pid 373, state active, sub running, last exit 0, reason 0) RPC probe: ok
Discovery (this machine) Found 0 gateways via Bonjour (local.) Tip: if the gateway is remote, mDNS won’t cross networks; use Wide-Area Bonjour (split DNS) or SSH tunnels.
Targets Local loopback ws://127.0.0.1:18789 Connect: failed – timeout
现在能想到的解决方案就是多次运行 openclaw doctor , 直到开始能正常响应,有的时候运行2次 就搞定了,有的时候运行3次,4次都没有搞定。这实际上就导致了低效率,因为你不能放心地给飞书机器人一个指令就可以不管它了,还得不断地去看web客户端的日志、会话、使用情况,确保它是在忙于工作暂时没有输出搭理我。如果看到日志没有变化了,使用情况 中的输出也没有变化了,往往意味着Agent没有响应了,这个时候就要判断说它不是忙于工作,而是“死机”了, 这个时候就要去多运行几次openclaw doctor ,直到把它重新激活。
有的时候会碰到这种情况: openclaw gateway status 显示running 状态
root@desktop-jacksen:# openclaw gateway status
Runtime: running (pid 29162, state active, sub running, last exit 0, reason 0) RPC probe: ok
Listening: 127.0.0.1:18789 Troubles: run openclaw status Troubleshooting: https://docs.openclaw.ai/troubleshooting root@desktop-jacksen:#
但是,web客户端已经没响应了:
这个时候连续运行多次 Openclaw doctor 也没有用。 再尝试运行openclaw gateway restart 依然没用。
root@desktop-jacksen:# openclaw gateway status Runtime: running (pid 29822, state active, sub running, last exit 0, reason 0) RPC probe: ok
Listening: 127.0.0.1:18789 Troubles: run openclaw status Troubleshooting: https://docs.openclaw.ai/troubleshooting root@desktop-jacksen:#
元宝说确实有守护进程:
openclaw daemon (守护进程,pid 29162)
└── openclaw gateway (网关子进程,pid 29822)
解决方案是停掉守护进程:
openclaw daemon stop
停掉之后再查gateway状态 :openclaw gateway status , 发现已经停止了:
Runtime: stopped (state inactive, sub dead, last exit 0, reason 1) RPC probe: failed RPC target: ws://127.0.0.1:18789 gateway closed (1006 abnormal closure (no close frame)): no close reason Gateway target: ws://127.0.0.1:18789 Source: cli –url Config: /root/.openclaw/openclaw.json
Service is loaded but not running (likely exited immediately). File logs: /tmp/openclaw/openclaw-2026-03-23.log Logs: journalctl –user -u openclaw-gateway.service -n 200 –no-pager
Troubles: run openclaw status Troubleshooting: https://docs.openclaw.ai/troubleshooting root@desktop-jacksen:#
但我还是想解决claude的问题,于是把这个问题给豆包,豆包说可以把限制代码注释掉,不过由于claude的源代码是压缩过的,不好注释,于是豆包就先备份,然后用sed命令替换掉那段检查代码(sed -i ‘s/console.error(“–dangerously-skip-permissions cannot be used with root\/sudo privileges for security reasons”),process.exit(1)/\/\/ bypass root check/g’ /usr/lib/node_modules/@anthropic-ai/claude-code/cli.js),然后执行发现出错了,这个错误其实是由于豆包替换的代码片段不全,新的代码其实比这个更多:
if(K===”bypassPermissions”||_){if(process.platform!==”win32″&typeof process.getuid===”function”&process.getuid()===0&process.env.IS_SANDBOX!==”1″&process.env.CLAUDE_CODE_BUBBLEWRAP!==”1″)console.error(“–dangerously-skip-permissions cannot be used with root/sudo privileges for security reasons”)
20:36:41 error [tools] read failed: ENOENT: no such file or directory, access ‘/root/.openclaw/skills/minimax-docx/Samples/DocumentCreationSamples.cs’ raw_params={“path”:”/root/.openclaw/skills/minimax-docx/Samples/DocumentCreationSamples.cs”,”limit”:100}
DONE Compiled successfully in 1392ms 3:15:24 PM File Size Gzipped dist/remoteEntry.js 5.5 KB 0.0 B dist/206.f765d364.async.js 685.0 B 0.0 B dist/p__hello-page.0c6c99dd.async.js 258.0 B 0.0 B dist/p__blank2.30ce5ca9.async.js 218.0 B 0.0 B Images and other types of assets omitted. Done in 7.43s.
root@desktop-jacksen:~/.openclaw/memory# sqlite3 main.sqlite SQLite version 3.37.2 2022-01-06 13:25:41 Enter “.help” for usage hints. sqlite> SELECT * FROM embedding_cache LIMIT 5; openai|text-embedding-v4|276e6f94020abed8088b5ce057c930ce515e022d4427e5271514cd8454496fe0|8130a66ade233348e7f6b27d82df8fe9f9671a1bee64b479bf08e653325e3a43|[-0.03580964356660843,0.001063261297531426,0.06607277691364288,-0.0539744570851326,0.020920734852552414,-0.025877926498651505,0.030592458322644234,0.003951887134462595,-0.018840795382857323,0.13082826137542725,0.009281734935939312,-0.008800748735666275,0.03813224285840988,0.04062817245721817,-0.019464777782559395,-0.002703922800719738,-0.006534480024129152,-0.057545021176338196,-0.0003791558265220374,-0.020608743652701378,-0.0369882769882679,-0.026207251474261284,0.03854823112487793,-0.022376693785190582,0.002420097589492798………
三、配置Agent团队
在OpenClaw的文档中有对多智能体的配置说明:
当前默认的Agent是H0-OpenClaw助手
创建H0开发助手
先看命令行创建Agent的帮助:
root@desktop-jacksen:~/.openclaw/memory# openclaw agents add –help chunks_ftschunks_ftschunks_ftschunks_ftschunks_fts �� OpenClaw 2026.2.24 (df9a474) — I run on caffeine, JSON5, and the audacity of “it worked on my machine.”
Usage: openclaw agents add [options] [name]
Add a new isolated agent
Options: –agent-dir <dir> Agent state directory for this agent –bind <channel[:accountId]> Route channel binding (repeatable) (default: []) -h, –help Display help for command –json Output JSON summary (default: false) –model <id> Model id for this agent –non-interactive Disable prompts; requires –workspace (default: false) –workspace <dir> Workspace directory for the new agent
org.springframework.http.converter.HttpMessageNotReadableException: JSON parse error: Cannot deserialize value of type `java.lang.Long` from String “instance-monitor”: not a valid `java.lang.Long` value
at org.springframework.http.converter.json.AbstractJackson2HttpMessageConverter.readJavaType(AbstractJackson2HttpMessageConverter.java:409) ~[spring-web-6.2.1.jar!/:6.2.1]
at org.springframework.http.converter.json.AbstractJackson2HttpMessageConverter.read(AbstractJackson2HttpMessageConverter.java:357) ~[spring-web-6.2.1.jar!/:6.2.1]
at org.springframework.web.servlet.mvc.method.annotation.AbstractMessageConverterMethodArgumentResolver.readWithMessageConverters(AbstractMessageConverterMethodArgumentResolver.java:204) ~[spring-webmvc-6.2.1.jar!/:6.2.1]
at org.springframework.web.servlet.mvc.method.annotation.RequestResponseBodyMethodProcessor.readWithMessageConverters(RequestResponseBodyMethodProcessor.java:176) ~[spring-webmvc-6.2.1.jar!/:6.2.1]
org.hzero.core.exception.ValidateFailureException: [collectionName must not be blank; ] at org.hzero.core.base.DefaultValidationResultProcessor.process(DefaultValidationResultProcessor.java:43) ~[hzero-starter-core-1.12.1.RELEASE.jar!/:1.12.1.RELEASE] at org.hzero.core.util.ValidUtils.valid(ValidUtils.java:59) ~[hzero-starter-core-1.12.1.RELEASE.jar!/:1.12.1.RELEASE] at org.hzero.core.base.BaseAppService.validObject(BaseAppService.java:35) ~[hzero-starter-core-1.12.1.RELEASE.jar!/:1.12.1.RELEASE] at org.hzero.aip.app.service.impl.VectorCollectionServiceImpl.create(VectorCollectionServiceImpl.java:95) ~[hzero-aip-server-1.6.1.ALPHA.6.13.jar!/:1.6.1.ALPHA.6.13]
java.lang.ClassCastException: org.hzero.aip.model.domain.vo.response.content.FailureResp incompatible with org.hzero.aip.model.domain.vo.response.content.chat.ChatCommonResp at org.hzero.aip.model.infra.helper.SyncResponseHelper.apply(SyncResponseHelper.java:98)
Hzero-aip-app服务中的错误信息:
2026-03-05 11:31:18.231 | org.springframework.web.client.HttpClientErrorException$BadRequest: 400 Bad Request on POST request for “https://ark.cn-beijing.volces.com/api/v3/chat/completions”: “{“error”:{“code”:”InvalidParameter”,”message”:”‘messages’ must contain the word ‘json’ in some form, to use ‘response_format’ of type ‘json_object’. Request id: 021772681472981c54cfbc8e773a7e1d131d5b1cad603e0f69745″,”param”:””,”type”:”BadRequest”}}”
删除业务对象时报错:”exception”: “Not found [PAGE_STRUCTURE], [B0] PivotProvider
Modeler服务后台错误是:io.choerodon.core.exception.ext.IllegalArgumentException: Not found [PAGE_STRUCTURE], [BO] PivotProvider.
at org.hzero.boot.modeler.pivot.PivotProviderManager.getPivotProvider(PivotProviderManager.java:269) ~[hzero-boot-modeler-client-2.10.2-1.12.BETA.4.jar!/:2.10.2-1.12.BETA.4]
===============
2026/01/19
仅针对在2026/01/19之前下载的朋友需替换文件(根据CPU类型选择):
1、镜像文件更新:无
2、数据文件更新:
./2026-01-19/x86/hzero-dockers.tar.gz
./2026-01-19/arm64/hzero-dockers-arm64.tar.gz
3、具体更新内容:
3.1、执行了update hzero_platform.iam_role set is_enabled=1 where id =5;解决新建租户失败的问题。
3.2、redis-docker目录下的docker-compose.yml增加配置:command: redis-server –save “” –appendonly no ; 目的是去掉了持久化,解决低资源环境下某些服务的日志中出现的redis 操作有时会的timeout问题。
2026-01-18 04:13:44.747 ERROR 1 — [ XNIO-1 task-5] o.j.o.s.a.s.i.ImportZipDataServiceImpl : MCP Get Data Mapping Has Error
feign.codec.DecodeException: Error while extracting response for type [java.util.List<org.jipaas.orchestration.service.app.dto.ImportMappingDTO>] and content type [application/json;charset=UTF-8]
at feign.InvocationContext.decode(InvocationContext.java:125)
……
Caused by: org.springframework.web.client.RestClientException: Error while extracting response for type [java.util.List<org.jipaas.orchestration.service.app.dto.ImportMappingDTO>] and content type [application/json;charset=UTF-8]
at org.springframework.web.client.HttpMessageConverterExtractor.extractData(HttpMessageConverterExtractor.java:127)
……
Caused by: org.springframework.http.converter.HttpMessageNotReadableException: JSON parse error: Cannot deserialize value of type `java.util.ArrayList<org.jipaas.orchestration.service.app.dto.ImportMappingDTO>` from Object value (token `JsonToken.START_OBJECT`)
……
Caused by: com.fasterxml.jackson.databind.exc.MismatchedInputException: Cannot deserialize value of type `java.util.ArrayList<org.jipaas.orchestration.service.app.dto.ImportMappingDTO>` from Object value (token `JsonToken.START_OBJECT`)
at [Source: REDACTED (`StreamReadFeature.INCLUDE_SOURCE_IN_LOCATION` disabled); line: 1, column: 1]
java.lang.NullPointerException: Cannot invoke “com.fasterxml.jackson.databind.JsonNode.asText()” because the return value of “com.fasterxml.jackson.databind.JsonNode.get(java.lang.String)” is null
ERROR in ./node_modules/hzero-front-hlod/lib/routes/NewRenderEngine/RenderContainerComponents/RenderSteps/RenderSteps.js 16:0-27
Module not found: Error: Package path ./v4 is not exported from package /d02/hzero/front/front/node_modules/hzero-front-haip-card/node_modules/uuid (see exports field in /d02/hzero/front/front/node_modules/hzero-front-haip-card/node_modules/uuid/package.json)
前端组说需要回滚:
那还是在 package.json中改回:
“hzero-front-haip-card”:”1.6.1-alpha.2.7″,
然后再yarn ,yarn run build:ms
yarn run build:ms hzero-front-haip,hzero-front-haip-card
Caused by: org.springframework.web.client.HttpClientErrorException$BadRequest: 400 Bad Request on POST request for “https://ark.cn-beijing.volces.com/api/v3/chat/completions”: “{“error”:{“code”:”InvalidParameter”,”message”:”schema must be provided. Request id: 021770360514613b1540e0519088842060f281dec849b042fa5ea”,”param”:””,”type”:”BadRequest”}}”
后台服务aip-app服务日志,显示是:2026-02-06T14:48:39.414+08:00 WARN 63182 — [hzero-aip-app] [ model-38] o.h.a.m.d.s.GradientIntervalStrategy : CHAT@HAIP.DOUBAO_CHAT_COPY is silenced, release time is: 2026-02-06T14:50:34.956188440
com.fasterxml.jackson.databind.exc.InvalidTypeIdException: Could not resolve subtype of [simple type, class org.hzero.aip.common.jsonschema.JsonSchema]: missing type id property ‘type’ (for POJO property ‘properties’)
有问题,看文件管理界面汇总查询也在打圈圈,看起来是minio服务出问题了,重启minio服务,重启后文件服务正常了,再测试编排流程:结果还是文件服务read timeOut了,查文件服务的日志,out of Memory了: 2026-02-09T20:10:34.012+08:00 ERROR 108664 — [hzero-file] [ XNIO-1 I/O-4] org.xnio.listener : XNIO001007: A channel event listener threw an exception
java.lang.OutOfMemoryError: Java heap space
at java.base/java.nio.ByteBuffer.allocateDirect(ByteBuffer.java:332) ~[na:na]
at io.undertow.server.DefaultByteBufferPool.allocate(DefaultByteBufferPool.java:149) ~[undertow-core-2.3.18.Final.jar!/:2.3.18.Final]
at io.undertow.server.protocol.http.HttpReadListener.handleEventWithNoRunningRequest(HttpReadListener.java:149) ~[undertow-core-2.3.18.Final.jar!/:2.3.18.Final]
at io.undertow.server.protocol.http.HttpReadListener.handleEvent(HttpReadListener.java:136) ~[undertow-core-2.3.18.Final.jar!/:2.3.18.Final]
### Cause: com.mysql.cj.jdbc.exceptions.MysqlDataTruncation: Data truncation: Data too long for column ‘document_name’ at row 1
; Data truncation: Data too long for column ‘document_name’ at row 1
at org.springframework.jdbc.support.SQLStateSQLExceptionTranslator.doTranslate(SQLStateSQLExceptionTranslator.java:118) ~[spring-jdbc-6.2.1.jar!/:6.2.1]
at org.springframework.jdbc.support.AbstractFallbackSQLExceptionTranslator.translate(AbstractFallbackSQLExceptionTranslator.java:107) ~[spring-jdbc-6.2.1.jar!/:6.2.1]
at org.springframework.jdbc.support.AbstractFallbackSQLExceptionTranslator.translate(AbstractFallbackSQLExceptionTranslator.java:116) ~[spring-jdbc-6.2.1.jar!/:6.2.1]
at org.hzero.core.util.ResponseUtils.lambda$static$1(ResponseUtils.java:54) ~[hzero-starter-core-1.12.1.RELEASE.jar!/:1.12.1.RELEASE]
at org.hzero.core.util.ResponseUtils.getResponse(ResponseUtils.java:230) ~[hzero-starter-core-1.12.1.RELEASE.jar!/:1.12.1.RELEASE]
at org.hzero.core.util.ResponseUtils.getResponse(ResponseUtils.java:86) ~[hzero-starter-core-1.12.1.RELEASE.jar!/:1.12.1.RELEASE]
at org.hzero.aip.agent.flow.execute.executor.impl.chat.ChatNodeHandler.createPersonalDocuments(ChatNodeHandler.java:1385) ~[hzero-aip-app-1.6.1.ALPHA.6.13.jar!/:1.6.1.ALPHA.6.13] 但是没有明确的错误信息,而且在aip-server中也没有出现错误。不知道是啥原因了:
### Cause: java.sql.SQLIntegrityConstraintViolationException: Column ‘field_type’ cannot be null
; Column ‘field_type’ cannot be null
at org.springframework.jdbc.support.SQLExceptionSubclassTranslator.doTranslate(SQLExceptionSubclassTranslator.java:97) ~[spring-jdbc-6.2.1.jar!/:6.2.1]
Caused by: org.springframework.beans.factory.UnsatisfiedDependencyException: Error creating bean with name ‘externalInvokeAdvisor’: Unsatisfied dependency expressed through field ‘advice’: Error creating bean with name ‘internalInvokeAdvisor’: Unsatisfied dependency expressed through field ‘advice’: Error creating bean with name ‘internalInvokeAdvice’: Unsatisfied dependency expressed through field ‘objectMapper’: Error creating bean with name ‘interfaceGatewayConfig’ defined in URL [jar:nested:/d02/hzero-ipaas/project/ps-jipaas-interface/target/ps-jipaas-interface.jar/!BOOT-INF/lib/hzero-interface-gateway-1.1.1.RELEASE.jar!/org/hzero/interfaces/gateway/config/InterfaceGatewayConfig.class]: Error creating bean with name ‘org.springframework.cache.annotation.ProxyCachingConfiguration’: No bean named ‘org.springframework.context.annotation.ConfigurationClassPostProcessor.importRegistry’ available
元宝分析原因:
主要原因与解决方案
这种内部 Bean 缺失通常不直接是业务代码错误,而是由以下情况引发的 Spring 容器初始化异常:
可能原因
说明与解决方案
版本冲突
这是最可疑的原因。不同 Spring 组件(如 spring-context, spring-aop, spring-cache)或第三方库(如 MyBatis)的版本不兼容,可能导致内部 Bean 注册紊乱 。解决方案:检查项目的依赖树(如使用 mvn dependency:tree或 gradle dependencies),确保所有 Spring 相关依赖的版本保持一致且兼容。
由于是笔记本上WSL虚拟机,我后来发现redis time out的问题不仅仅是这个持久化写的频率问题, 笔记本的电源计划在某天升级Windows后变成了节能模式,在接通电源情况3分钟就休眠,这也是导致各服务卡顿的主要问题。后来就更改了电源计划,在接通电源的模式下改成永不休眠。 卡顿的问题就得到了改善。
org.springframework.beans.factory.BeanCurrentlyInCreationException: Error creating bean with name ‘observationRestTemplateCustomizer’: Requested bean is currently in creation: Is there an unresolvable circular reference or an asynchronous initialization dependency?
at org.springframework.beans.factory.support.DefaultSingletonBeanRegistry.beforeSingletonCreation(DefaultSingletonBeanRegistry.java:421) ~[spring-beans-6.2.1.jar!/:6.2.1]
at org.springframework.beans.factory.support.DefaultSingletonBeanRegistry.getSingleton(DefaultSingletonBeanRegistry.java:282) ~[spring-beans-6.2.1.jar!/:6.2.1]
at org.springframework.beans.factory.support.AbstractBeanFactory.doGetBean(AbstractBeanFactory.java:334) ~[spring-beans-6.2.1.jar!/:6.2.1]
元宝分析:问题根源:循环依赖
错误信息 BeanCurrentlyInCreationException和 Is there an unresolvable circular reference?明确指出了问题的核心:Spring 容器在初始化 Bean 时,发现它们之间存在直接的或间接的循环引用关系 。
at org.hzero.workflow.engine.run.impl.ModelServiceImpl.checkValid(ModelServiceImpl.java:656) ~[hzero-workflow-1.20.0-1.12.RELEASE.jar!/:1.20.0-1.12.RELEASE]
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method) ~[na:na]
Caused by: org.springframework.beans.factory.BeanCreationException: Error creating bean with name ‘org.springframework.cache.annotation.ProxyCachingConfiguration’: No bean named ‘org.springframework.context.annotation.ConfigurationClassPostProcessor.importRegistry’ available