All Shares appear to be missing [ReadyNAS Ultra 2; Radiator 4.2.28]

[TigerRob]

Some weeks ago I (possibly stupidly) ignored a reported filesystem consistency check failure.  Have had the check pass successfully since.

Today I discovered that I can no longer connect to any of my shares (via SMB) from any Windows 10, Mac, or iOS devices.  I can access the one share that I have made available via HTTP.  I can access the Admin UI.

I went to have a look at the logs, and that screen says 'No logs available'.  (I do have a subset of logs available form a remote syslog server).  This line looked interesting 🙂

Nov 15 20:47:34 TMF-UK-PL-STORE01 syslogd:  /var/log/user.log: No space left on device

SSHing in it appears that /dev/md0 is out of space:

Store01:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/md0               4176564   4176564         0 100% /
tmpfs                       16         0        16   0% /USB
/dev/c/c             2903445768 375335472 2528110296  13% /c
Store01:~#

Help with how to proceed from here??

Thanks in advance,

Rob.

[TigerRob · Accepted solution]

Thank you all for your help (esp. @StephenB for the mention of crashplan)

3.5GB in /usr/local/crashplan/iupgrade ...

Stopped crashplan, moved that directory off to the data partition and softlinked to it, rebooted, and I appear to be back in business.

[bedlam1]

This post might help you out as a start https://community.netgear.com/t5/Using-your-ReadyNAS/How-do-i-clean-md0/td-p/1138772

One other option is to perform a Factory Reset (this wipes ALL data, so get a full backup), if you go down that route you may want to consider upgrading to the latest OS 6 system

see here http://netgear.nas-central.org/wiki/OS6_on_Legacy_x86

[StephenB]

Do you have crashplan installed?

[TigerRob]

Thanks @bedlam1.  Some good pointers to get me going.  Will start analysing tonight.   Whilst I have backups and factory reset is an option, I hope that won't be necessary.

[TigerRob]

@StephenB - yes I am running Crashplan.  From other threads it appears there might be some Crashplan directories that grow (and should be moved onto a data partition).  Pointers for cleaning up the crashplan installation and moving the appropriate directory would be much appreciated.

[bedlam1]

@TigerRob you are welcome, installing OS 6 will give you a whole lot more functionality though, you will have to reinstall apps though, and other settings..

It works like a charm on my legacy Readynas Pro 4 

[TigerRob]

@StephenB Found this thread https://community.netgear.com/t5/Backing-up-to-your-ReadyNAS/Crashplan-JRE-update-needed/m-p/987406#...

I suspect that gives me enough pointers - will investigate later.

[TigerRob · Accepted solution]

Thank you all for your help (esp. @StephenB for the mention of crashplan)

3.5GB in /usr/local/crashplan/iupgrade ...

Stopped crashplan, moved that directory off to the data partition and softlinked to it, rebooted, and I appear to be back in business.

[StephenB]

You will need to get the upgrade to 4.8 to work, otherwise the problem will come back.  I had to create a .wgetrc file to disable the certificate check to get the upgrade to complete.

[TigerRob]

@StephenB thanks for the heads up.  This thread is now well OT, but ...

Crashplan apparently didn't like me soft-linking the upgrade directory over to the data partition, so I re-instated a real (but empty) upgrade directory.

Then saw what you meant by the wget problem, and so added a check_certificate=off directive to /etc/wgetrc.  That solved the download issue, and the upgrade proceeded.  LOkking at the logs it didn't complete correctly, but appears that the failure was somewhere towards the end of the upgrade script where it was attempting to upgrade teh GUI, and we don't need that ..

But my 4.8.0 install fails with errors in the engine_errors.log - it appears to be this issue https://www.reddit.com/r/synology/comments/550806/best_crashplan_docker_image/d87beo9/ and modifying FULL_CP, as suggested gets me to a different error, which appears to be a 32- vs. 64-bit libstdc++ error:

java.lang.UnsatisfiedLinkError: libstdc++.so.6: wrong ELF class: ELFCLASS32

More research needed to find out how to get a 64-bit version installed and get Crashplan to find it ...

[TigerRob]

I'm wondering whether "downgrading" to a 32-bit JVM might resolve this.

@StephenB do you have Crashplan 4.8.0 running successfully on you readyNAS?  If yes, then which JVM?

[StephenB]

---

@TigerRob wrote:

I'm wondering whether "downgrading" to a 32-bit JVM might resolve this.

@StephenB do you have Crashplan 4.8.0 running successfully on you readyNAS?  If yes, then which JVM?

---

4.8 is running on my pro, using the java that Crashplan installed.   Per install.vars

JAVACOMMON=/usr/local/crashplan/jre/bin/java

The library path in apps.log is

LibraryPath = /usr/java/packages/lib/i386:/lib:/usr/lib:.:lib

checking the version gives me this:

PRO:/usr/local/crashplan/jre/bin# ./java -version

java version "1.8.0_72"
Java(TM) SE Runtime Environment (build 1.8.0_72-b15)
Java HotSpot(TM) Server VM (build 25.72-b15, mixed mode)

[TigerRob]

Why can't I post to this thread?

[TigerRob]

So, I have a 64-bit JVM.   Replacing it with a 32-bit JVM has made my crashplan install work again.

I'm having trouble posting the detail to this thread, my post keeps disappearing.  If this is successful then I'll add details.

[TigerRob]

Not sure why I'm having trouble posting the details to this thread.  I'm now removing all formatting (although I wasn;t using much before) and hope it'll get through this time.

So, I have a 64-bit Java installation.  On investigation the crashplan install script uses "uname -m" to determine which JVM to install, and so on a ReadyNAS will pick a 64-bit bundled JVM.   On upgrade the upgrade scripts looks at the installed JVM (whereever JAVACOMMON points) to determine whether to use a 32- or 64-bit version.  So, if you installed a 32-bit Java before Crashplan insisted on using the bundled JVM then your upgrades will have stuck to a 32-bit version.  For me it weas keeping a 64-bit version, as that was what was bundled when I did my original install.

I simply replaced /usr/local/crashplan/jre with the 32-bit JVM downl;oaded from code42 (see upgrade.properties for the URL) and I have a happy, working Crashplan again.

Thanks all,

Rob.

[StephenB]

---

@TigerRob wrote:

So, I have a 64-bit Java installation.  On investigation the crashplan install script uses "uname -m" to determine which JVM to install, and so on a ReadyNAS will pick a 64-bit bundled JVM.   On upgrade the upgrade scripts looks at the installed JVM (whereever JAVACOMMON points) to determine whether to use a 32- or 64-bit version.  So, if you installed a 32-bit Java before Crashplan insisted on using the bundled JVM then your upgrades will have stuck to a 32-bit version.  For me it weas keeping a 64-bit version, as that was what was bundled when I did my original install.

---

This explains why some users end up with 32 JVMs (with Crashplan working) while others ended up with 64-bit.  Thanks for sharing.

[StephenB]

---

@TigerRob wrote:

Not sure why I'm having trouble posting the details to this thread.

---

Something was triggering the automatic spam filter - not sure what.