Working with Compressed (.gz) Statistics Archive Files
search cancel

Working with Compressed (.gz) Statistics Archive Files

book

Article ID: 294076

calendar_today

Updated On:

Products

VMware Tanzu Gemfire

Issue/Introduction

Symptoms:

Currently, Visual Statistics Display (VSD) can only load uncompressed statistic archives files. So, automatically compressed statistics archive files (i.e. those where the configured archive file name has a ".gz" suffix) must first be uncompressed using a tool like gunzip. However, if the gfs.gz file was not properly closed, which happens if the server writing it is still running or if the server was killed or crashed, the gunzip command may fail. Options for extracting such files are discussed here.

Using the gunzip command on automatically compressed statistics archive files fails with an "unexpected end of file" error.

Environment


Cause

GemFire statistics are written as an "outputStream" to the statistics archive file while gzip is a block compression algorithm. Hence, if the file is not properly padded and closed out, the final block will not be a "proper" gzip block.  This causes problems for some tools, such as gunzip.

Resolution

Our recommendation is simply to avoid using the compression feature by setting the statistic-archive-file to a value without the ".gz" suffix. However, if you need this feature and/or already have compressed statistics files, you can convert them into a usable, uncompressed format using gzcat and redirecting the output to a new file, i.e.:

$ gzcat statistics.gfs.gz > statistics.gfs

or, equivalently, with gzip:

$ gzip -d -c statistics.gfs.gz > statistics.gfs

This method will still give a message about an unexpected EOF but it can be ignored and the resulting "statistics.gfs" should successfully load into VSD.