NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Solved
Out of memory isra 13+31c and RNDP4000 Lock up after upgrade to OS6 from OS4
after an initial successful upgrade from OS 4 to 6.1 and 4 x 8TB drives we've begun to experience access issues with our Readynas Pro 4 (RN). The error is repeatable as per the five steps outlined below (although we're not exactly sure which steps/combination of steps restores access) and the only workaround that will alleviate our issue is to;
- first login to FRONTVIEW
- select the SHARES panel
- EXPLORE a few of the shares and then
- click on REFRESH.
this has the effect of 'unblocking' whatever is causing the RN to lock up and then the unit functions normally thereafter.
It appears to be an issue with the WRITE function as the shares can be accessed and READ via Frontview / Explorer without any issues at first boot but as soon as something needs to be written to the unit it locks up.
on quite a few occasions after the unit locks up we’ve had the Out of memory isra 13+31c message pop up on the LED screen
and once the LED showed Network Changed.
system settings;
OS: 6.6.1-T230
Services: SMB, UPnP, HTTP, HTTPS & SMB Plus (Protocol 2.0)
Share Services: SMB Only, Anti-virus / Bit Rot are OFF & Snapshots OFF*
*there are NO Snapshots on the RN as these were all disabled when we first created our shares.
what causes the error;
- Boot the RN
- Access the RN (e.g. either via our file synchronising program Goodsync (GS) or by running a maintenance command such as Defrag))
- After GS ANALYSE has completed we click SYNC and files begin copying (70/80 MB/s) / Defrag starts for 5-10 seconds with the activity light illuminated before RN slows down to a halt
- GS reports a network error (cannot connect), Defrag stops and any EXPLORER window previously open which had been accessing shares now reports ‘Windows cannot access \\**’
- RN is now frozen and unresponsive (LED won't light up either) and only holding down the power button / pulling the cable to force a hard shutdown enables us to reboot the RN
the migration to OS6 has generally been pretty painless and there's lots to love about it. However the locking up is a major issue for us and (judging by the amount of forum posts we've read about this very problem) it's clearly a considerable stumbling block for many others.
It would be brilliant if we can put this issue to bed so again any help would be very much appreciated.
thanks
Have you considered upgrading the memory on the Pro?
11 Replies
Replies have been turned off for this discussion
You shouldn't stick on old beta firmware.
Update to 6.7.1 and then perhaps disable quotas on the volume.
thanks for your suggestion.
we're a little hesitant about upgrading to 6.7.1 due to not being able to downgrade, the numerous posts
of other users under 6.7.1 having lockups and our own experience of the 6.6.1 to 6.6.1 t230 upgrade
failing to solve our issue, but having read this, we looked at the metadata setting under our own
btrfs.log and low and behold we too have a huge amount of metadata.=== filesystem /data ===
Data, single: total=10.59TiB, used=10.50TiB
System, DUP: total=8.00MiB, used=1.12MiB
Metadata, DUP: total=5.50GiB, used=4.55GiB
GlobalReserve, single: total=512.00MiB, used=0.00B
with a very simplistic understanding of the RN's mechanics it seems completely plausible to link
oversized metadata content with the out of memory messages we've been getting so perhaps there's
hope yet! we understand that we'd need to use SSH and # btrfs quota disable /data to disable quotas
but is there any way (without having to carry out a third factory reset) to reduce the current size of
our metadata or is it only through carrying out the steps of
1. a factory reset
2. disabling quotas and then
3. restoring fresh a backup
the only way to keep the metadata size down?For the amount of data you have the amount of metadata you have is quite reasonable.
Disabling/enabling quotas would have minimal effect on the amount of metadata. Presumably quotas would be stored in metadata, but it doesn't take much space to store a few numbers.
metadata is data about data. Lots of CoW (Copy on Write) leads to a lot of metadata.
6.7.1 has a GUI option to disable quotas on the volume.
Disabling quotas on the volume is a good troubleshooting step. If that doesn't resolve it then there are others reasons for OOMing that could be considered. If you review the logs you should see which processes were using lots of memory when it OOMed.
It's been a good few months since we last posted which has seen our RN sat in the corner gathering dust primarily due to the locking up issue and losing complete interest/adopting an alternative backup strategy. We approached the RN again today as a last ditched attempt thinking we'd try updating to 6.8.1 and if that doesn't help then we'll reset and go back to OS4. Sadly despite the upgrade to 6.8.1, disabling quotas along with checksums, this has done nothing to solve our problem. We'd prefer not to spend time fiddling around trying to get this to work but before we reset with OS4 and begin copy everything over again, are there any firm suggestions, sure-fire fixes or specific logs to examne that others have had where they've experienced the locking up/out of memory issue BUT actually managed to fix it?
ThanksHave you considered upgrading the memory on the Pro?
We did read up on a possible CPU/Memory upgrade a while back and it appeared that a Core 2 E7600 cpu and memory with timings of 5-5-5-15 were recommended. Is that still the case and besides the Crucial CT2KIT25664AA1067 (PC3-8500 1066MHz Memory) which popped up a few times what memory would you suggest please?
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
