The HTTP/2 implementation in Apache Tomcat 9.0.0.M1 to 9.0.14 and 8.5.0 to 8.5.37 accepted streams with excessive numbers of SETTINGS frames and also permitted clients to keep streams open without reading/writing request/response data. By keeping streams open for requests that utilised the Servlet API's blocking I/O, clients were able to cause server-side threads to block eventually leading to thread exhaustion and a DoS.

CXF supports (via JwtRequestCodeFilter) passing OAuth 2 parameters via a JWT token as opposed to query parameters (see: The OAuth 2.0 Authorization Framework: JWT Secured Authorization Request (JAR)). Instead of sending a JWT token as a "request" parameter, the spec also supports specifying a URI from which to retrieve a JWT token from via the "request_uri" parameter. CXF was not validating the "request_uri" parameter (apart from ensuring it uses "https) and was making a REST request to the parameter in the request to retrieve a token. This means that CXF was vulnerable to DDos attacks on the authorization server, as specified in section 10.4.1 of the spec. This issue affects Apache CXF versions prior to 3.4.3; Apache CXF versions prior to 3.3.10.

When there are multiple ranges in a range request, Apache Traffic Server (ATS) will read the entire object from cache. This can cause performance problems with large objects in cache. This affects versions 6.0.0 to 6.2.2 and 7.0.0 to 7.1.3. To resolve this issue users running 6.x users should upgrade to 6.2.3 or later versions and 7.x users should upgrade to 7.1.4 or later versions.

Apache ATS 6.0.0 to 6.2.3, 7.0.0 to 7.1.9, and 8.0.0 to 8.0.6 is vulnerable to a HTTP/2 slow read attack.

The documentation of Apache Tomcat 10.1.0-M1 to 10.1.0-M14, 10.0.0-M1 to 10.0.20, 9.0.13 to 9.0.62 and 8.5.38 to 8.5.78 for the EncryptInterceptor incorrectly stated it enabled Tomcat clustering to run over an untrusted network. This was not correct. While the EncryptInterceptor does provide confidentiality and integrity protection, it does not protect against all risks associated with running over any untrusted network, particularly DoS risks.

By specially crafting HTTP/2 requests, workers would be allocated 60 seconds longer than necessary, leading to worker exhaustion and a denial of service. Fixed in Apache HTTP Server 2.4.34 (Affected 2.4.18-2.4.30,2.4.33).

In Apache HTTP server versions 2.4.37 and prior, by sending request bodies in a slow loris way to plain resources, the h2 stream for that request unnecessarily occupied a server thread cleaning up that incoming data. This affects only HTTP/2 (mod_http2) connections.

A vulnerability in the JsonMapObjectReaderWriter of Apache CXF allows an attacker to submit malformed JSON to a web service, which results in the thread getting stuck in an infinite loop, consuming CPU indefinitely. This issue affects Apache CXF versions prior to 3.4.4; Apache CXF versions prior to 3.3.11.

Two four letter word commands "wchp/wchc" are CPU intensive and could cause spike of CPU utilization on Apache ZooKeeper server if abused, which leads to the server unable to serve legitimate client requests. Apache ZooKeeper thru version 3.4.9 and 3.5.2 suffer from this issue, fixed in 3.4.10, 3.5.3, and later.

In Apache Thrift 0.9.3 to 0.13.0, malicious RPC clients could send short messages which would result in a large memory allocation, potentially leading to denial of service.

Some HTTP/2 implementations are vulnerable to a settings flood, potentially leading to a denial of service. The attacker sends a stream of SETTINGS frames to the peer. Since the RFC requires that the peer reply with one acknowledgement per SETTINGS frame, an empty SETTINGS frame is almost equivalent in behavior to a ping. Depending on how efficiently this data is queued, this can consume excess CPU, memory, or both.

Some HTTP/2 implementations are vulnerable to a reset flood, potentially leading to a denial of service. The attacker opens a number of streams and sends an invalid request over each stream that should solicit a stream of RST_STREAM frames from the peer. Depending on how the peer queues the RST_STREAM frames, this can consume excess memory, CPU, or both.

Some HTTP/2 implementations are vulnerable to ping floods, potentially leading to a denial of service. The attacker sends continual pings to an HTTP/2 peer, causing the peer to build an internal queue of responses. Depending on how efficiently this data is queued, this can consume excess CPU, memory, or both.

Some HTTP/2 implementations are vulnerable to resource loops, potentially leading to a denial of service. The attacker creates multiple request streams and continually shuffles the priority of the streams in a way that causes substantial churn to the priority tree. This can consume excess CPU.

Some HTTP/2 implementations are vulnerable to a flood of empty frames, potentially leading to a denial of service. The attacker sends a stream of frames with an empty payload and without the end-of-stream flag. These frames can be DATA, HEADERS, CONTINUATION and/or PUSH_PROMISE. The peer spends time processing each frame disproportionate to attack bandwidth. This can consume excess CPU.

Some HTTP/2 implementations are vulnerable to a header leak, potentially leading to a denial of service. The attacker sends a stream of headers with a 0-length header name and 0-length header value, optionally Huffman encoded into 1-byte or greater headers. Some implementations allocate memory for these headers and keep the allocation alive until the session dies. This can consume excess memory.

Some HTTP/2 implementations are vulnerable to window size manipulation and stream prioritization manipulation, potentially leading to a denial of service. The attacker requests a large amount of data from a specified resource over multiple streams. They manipulate window size and stream priority to force the server to queue the data in 1-byte chunks. Depending on how efficiently this data is queued, this can consume excess CPU, memory, or both.

Some HTTP/2 implementations are vulnerable to unconstrained interal data buffering, potentially leading to a denial of service. The attacker opens the HTTP/2 window so the peer can send without constraint; however, they leave the TCP window closed so the peer cannot actually write (many of) the bytes on the wire. The attacker then sends a stream of requests for a large response object. Depending on how the servers queue the responses, this can consume excess memory, CPU, or both.

In Apache Qpid Broker-J versions 6.1.0 through 6.1.4 (inclusive) the broker does not properly enforce a maximum frame size in AMQP 1.0 frames. A remote unauthenticated attacker could exploit this to cause the broker to exhaust all available memory and eventually terminate. Older AMQP protocols are not affected.

The Security Team noticed that the termination condition of the for loop in the readExternal method is a controllable variable, which, if tampered with, may lead to CPU exhaustion. As a fix, we added an upper bound and termination condition in the read and write logic. We classify it as a "low-priority but useful improvement". SystemDS is a distributed system and needs to serialize/deserialize data but in many code paths (e.g., on Spark broadcast/shuffle or writing to sequence files) the byte stream is anyway protected by additional CRC fingerprints. In this particular case though, the number of decoders is upper-bounded by twice the number of columns, which means an attacker would need to modify two entries in the byte stream in a consistent manner. By adding these checks robustness was strictly improved with almost zero overhead. These code changes are available in versions higher than 2.2.1.

The Traffic Router component of the incubating Apache Traffic Control project is vulnerable to a Slowloris style Denial of Service attack. TCP connections made on the configured DNS port will remain in the ESTABLISHED state until the client explicitly closes the connection or Traffic Router is restarted. If connections remain in the ESTABLISHED state indefinitely and accumulate in number to match the size of the thread pool dedicated to processing DNS requests, the thread pool becomes exhausted. Once the thread pool is exhausted, Traffic Router is unable to service any DNS request, regardless of transport protocol.

In Apache SpamAssassin before 3.4.3, a message can be crafted in a way to use excessive resources. Upgrading to SA 3.4.3 as soon as possible is the recommended fix but details will not be shared publicly.

qpid-cpp: ACL policies only loaded if the acl-file option specified enabling DoS by consuming all available file descriptors

The ASN.1 parser in Bouncy Castle Crypto (aka BC Java) 1.63 can trigger a large attempted memory allocation, and resultant OutOfMemoryError error, via crafted ASN.1 data. This is fixed in 1.64.

The AsyncResponseWrapperImpl class in Apache Olingo versions 4.0.0 to 4.6.0 reads the Retry-After header and passes it to the Thread.sleep() method without any check. If a malicious server returns a huge value in the header, then it can help to implement a DoS attack.

Apache POI in versions prior to release 3.17 are vulnerable to Denial of Service Attacks: 1) Infinite Loops while parsing crafted WMF, EMF, MSG and macros (POI bugs 61338 and 61294), and 2) Out of Memory Exceptions while parsing crafted DOC, PPT and XLS (POI bugs 52372 and 61295).

A Denial of Service vulnerability was found in Apache Qpid Broker-J versions 6.0.0-7.0.6 (inclusive) and 7.1.0 which allows an unauthenticated attacker to crash the broker instance by sending specially crafted commands using AMQP protocol versions below 1.0 (AMQP 0-8, 0-9, 0-91 and 0-10). Users of Apache Qpid Broker-J versions 6.0.0-7.0.6 (inclusive) and 7.1.0 utilizing AMQP protocols 0-8, 0-9, 0-91, 0-10 must upgrade to Qpid Broker-J versions 7.0.7 or 7.1.1 or later.

An issue was discovered in the protobuf crate before 2.6.0 for Rust. Attackers can exhaust all memory via Vec::reserve calls.

It was discovered that the C++ implementation (which underlies the R, Python and Ruby implementations) of Apache Arrow 0.14.0 to 0.14.1 had a uninitialized memory bug when building arrays with null values in some cases. This can lead to uninitialized memory being unintentionally shared if Arrow Arrays are transmitted over the wire (for instance with Flight) or persisted in the streaming IPC and file formats.

The file name encoding algorithm used internally in Apache Commons Compress 1.15 to 1.18 can get into an infinite loop when faced with specially crafted inputs. This can lead to a denial of service attack if an attacker can choose the file names inside of an archive created by Compress.

The mod_http2 module in the Apache HTTP Server 2.4.17 through 2.4.23, when the Protocols configuration includes h2 or h2c, does not restrict request-header length, which allows remote attackers to cause a denial of service (memory consumption) via crafted CONTINUATION frames in an HTTP/2 request.

HTTP/2 (2.4.20 through 2.4.39) very early pushes, for example configured with "H2PushResource", could lead to an overwrite of memory in the pushing request's pool, leading to crashes. The memory copied is that of the configured push link header values, not data supplied by the client.

Apache Traffic Server is vulnerable to HTTP/2 setting flood attacks. Earlier versions of Apache Traffic Server didn't limit the number of setting frames sent from the client using the HTTP/2 protocol. Users should upgrade to Apache Traffic Server 7.1.7, 8.0.4, or later versions.

The fix for CVE-2019-0199 was incomplete and did not address HTTP/2 connection window exhaustion on write in Apache Tomcat versions 9.0.0.M1 to 9.0.19 and 8.5.0 to 8.5.40 . By not sending WINDOW_UPDATE messages for the connection window (stream 0) clients were able to cause server-side threads to block eventually leading to thread exhaustion and a DoS.

An access permission override in Apache Struts 2.0.0 to 2.5.20 may cause a Denial of Service when performing a file upload.

In Apache Subversion versions up to and including 1.9.10, 1.10.4, 1.12.0, Subversion's svnserve server process may exit when a client sends certain sequences of protocol commands. This can lead to disruption for users of the server.

A bug exists in the way mod_ssl handled client renegotiations. A remote attacker could send a carefully crafted request that would cause mod_ssl to enter a loop leading to a denial of service. This bug can be only triggered with Apache HTTP Server version 2.4.37 when using OpenSSL version 1.1.1 or later, due to an interaction in changes to handling of renegotiation attempts.

In Apache ActiveMQ 5.0.0 - 5.15.8, unmarshalling corrupt MQTT frame can lead to broker Out of Memory exception making it unresponsive.

A vulnerability was found in Apache HTTP Server 2.4.17 to 2.4.38. Using fuzzed network input, the http/2 request handling could be made to access freed memory in string comparison when determining the method of a request and thus process the request incorrectly.

In Apache Thrift 0.9.3 to 0.12.0, a server implemented in Go using TJSONProtocol or TSimpleJSONProtocol may panic when feed with invalid input data.

The HTTP/2 header parser in Apache Tomcat 9.0.0.M1 to 9.0.0.M11 and 8.5.0 to 8.5.6 entered an infinite loop if a header was received that was larger than the available buffer. This made a denial of service attack possible.

Apache HTTP Server versions 2.4.0 to 2.4.46 A specially crafted Cookie header handled by mod_session can cause a NULL pointer dereference and crash, leading to a possible Denial Of Service

A DNS proxy and possible amplification attack vulnerability in WebClientInfo of Apache Wicket allows an attacker to trigger arbitrary DNS lookups from the server when the X-Forwarded-For header is not properly sanitized. This DNS lookup can be engineered to overload an internal DNS server or to slow down request processing of the Apache Wicket application causing a possible denial of service on either the internal infrastructure or the web application itself. This issue affects Apache Wicket Apache Wicket 9.x version 9.2.0 and prior versions; Apache Wicket 8.x version 8.11.0 and prior versions; Apache Wicket 7.x version 7.17.0 and prior versions and Apache Wicket 6.x version 6.2.0 and later versions.

A carefully crafted invalid TLS handshake can cause Apache Traffic Server (ATS) to segfault. This affects version 6.2.2. To resolve this issue users running 6.2.2 should upgrade to 6.2.3 or later versions.

In Apache ORC 1.0.0 to 1.4.3 a malformed ORC file can trigger an endlessly recursive function call in the C++ or Java parser. The impact of this bug is most likely denial-of-service against software that uses the ORC file parser. With the C++ parser, the stack overflow might possibly corrupt the stack.

Unknown vulnerability in filestat.c for Apache running on OS2, versions 2.0 through 2.0.45, allows unknown attackers to cause a denial of service via requests related to device names.

The Catalina org.apache.catalina.connector.http package in Tomcat 4.0.x up to 4.0.3 allows remote attackers to cause a denial of service via several requests that do not follow the HTTP protocol, which causes Tomcat to reject later requests.

The ap_proxy_http_process_response function in mod_proxy_http.c in the mod_proxy module in the Apache HTTP Server 2.0.63 and 2.2.8 does not limit the number of forwarded interim responses, which allows remote HTTP servers to cause a denial of service (memory consumption) via a large number of interim responses.