#6 Failure, wrapped MIRA process aborted. While writing normalized skim block

[Mikhail]

I am trying to run test transcriptome assembly on MIRA with about 40 million Illumina reads using a manifest file like this

---

project = test
job = est,denovo,accurate
parameters = -GE:not=8 \
-HS:mnr=yes:nrr=15
readgroup = defaultGroup
segmentplacement = ---> <---
segmentnaming = solexa
data = SRR536882.solexa.fastq
technology = solexa

---

I am getting an error like this

---

...

Total megahubs: 0
Cutting back possible chimeras ... done.
Chimeras were searched for ... looking for hits to purge.
truncating test_assembly/test_d_tmp/test_int_posmatchf_pass.1.bin from 13020722856 to 13018027032
truncating test_assembly/test_d_tmp/test_int_posmatchc_pass.1.bin from 10357171608 to 10355046672
System memory: 101366665216
Mem2keepfree: 15204999782
Used by MIRA: 39189258240
Mem avail: 46972407194
rsh increased memtouse to: 46972407194
Edge vector capacity: 563719390
Can load up to 563714390 skim edges at once.
Partitioning into 4 blocks.
Blocks: 10655141 20968821 31383732 42052242
We have 1947756142 skims in file.
Localtime: Mon Nov 11 18:06:20 2013
De-normalising SKIM hits ... (this will take a while)
Localtime: Mon Nov 11 18:07:39 2013
Localtime: Mon Nov 11 18:09:15 2013
Writing normalised skimblock 0 ( 16.0 GiB) ... done.
Localtime: Mon Nov 11 18:16:06 2013
Localtime: Mon Nov 11 18:17:41 2013
Writing normalised skimblock 10655141 ( 16.0 GiB) ... Failure, wrapped MIRA process aborted.

[Bastien Chevreux]

The main MIRA process stopped without further message? This is very uncommon and I see only two possibilities for this: an error in MIRA leading to a memory fault ... or the OS killed MIRA (for whatever reason).

However, in both cases there should be additional text written to the stderr, but I don't see that in your pasted excerpt. Am I correct in guessing you redirected the output with ">" instead of ">&" (or "2>" if you're using simple shell)? If yes, can you please rerun and also capture stderr? Thanks.

[Mikhail]

The issue was: OS killed MIRA without printing anything to SDTERR because there was not enough disk space : ). Problem solved. Thanks.

[Bastien Chevreux]

• Description has changed:

Diff:

--- old
+++ new
@@ -1,4 +1,3 @@
-

 I am trying to run test transcriptome assembly on MIRA with about 40 million Illumina reads using a manifest file like this

• status: open --> closed