Article ID: 122532, created on Jul 31, 2014, last review on Apr 21, 2015

  • Applies to:
  • Virtuozzo 6.0

Symptoms

CS stopped to function properly with the following messages in the log file (/var/log/pstorage/$clustername/cs-xcxu63/cs.log.gz):

31-07-14 10:55:05.966 csd_send_io_reply: [2.18606:2958921761] IO error 258 29466624+65536@O0001d2b6 err=6
31-07-14 10:55:05.967 csd_input_iomsg: write on error chunk
31-07-14 10:55:12.372 csd_input_iomsg: write on error chunk
31-07-14 10:55:12.424 csd_input_iomsg: write on error chunk
31-07-14 10:55:12.489 csd_input_iomsg: write on error chunk
31-07-14 10:55:12.553 csd_input_iomsg: write on error chunk
31-07-14 10:55:12.608 csd_input_iomsg: write on error chunk

pstorage top may show the corresponding CS switch between active/inactive states or be marked as 'failed' eventually.

At the same time, the server might have high LA values and high memory usage due to keeping large amount of data in file caches.

Cause

The process is no longer able to create and update files located in chunk or journal repositories, e.g. there is no proper permission to the directories/files, or the file system has been remounted in read-only mode due to errors in its journal area.

Resolution

Check if this is related to broken hard disk (replace it if needed), and recreate the failed chunk server in accordance with the instructions provided in the documentation, Parallels Cloud Storage Administrator's Guide:

Search Words

io load

Strange problem with chunks

c62e8726973f80975db0531f1ed5c6a2 2897d76d56d2010f4e3a28f864d69223 0dd5b9380c7d4884d77587f3eb0fa8ef

Email subscription for changes to this article
Save as PDF