# Internal Threads Error - pigz

**URL:** https://forums.suse.com/t/internal-threads-error-pigz/33748
**Category:** SLED Hardware
**Created:** [April 2, 2019, 10:45am UTC](https://forums.suse.com/t/internal-threads-error-pigz/33748 "2019-04-02T10:45:57Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![Mithras](https://avatars.discourse-cdn.com/v4/letter/m/f04885/32.png) [@Mithras](https://forums.suse.com/u/Mithras)
#### Post date: [April 2, 2019, 10:45am UTC](https://forums.suse.com/t/internal-threads-error-pigz/33748/1 "2019-04-02T10:45:57Z")

</div>

Hey,

I am expeting a very strange behavior on an HP ProLiant DL380 Gen10. I’m not sure if it is hardware or software related.

When I use pigz (SLES default package) and specify the number of threads (4 in this example)

```auto
zcat AmP1_R1.fastq.gz | pigz -p 4 > out.fastq.gz
```

I get this error (5 out of 10 times):

```auto
pigz: abort: internal threads error
```

Strace says:

[QUOTE]write(1, “T  
TCACCGTGCCAGACTAGAGTCAAGCTCAAC”…, 32644) = 32644  
read(10, “\254Ml@\344\300\351;\216%\356\v\253Ce\26Q\306\250\324\265\f\350\312\362\301 #F\353\313\341”…, 32768) = 32768  
pigz: abort: internal threads error  
write(1, “@NB551369:35:H3VGYBGX9:2:21304:1”…, 32768) = 32768  
write(1, “ATGCCTCGTCATCTAATTAGTGACGCGCATGA”…, 32638) = 32638  
write(1, “@NB551369:35:H3VGYBGX9:2:21304:1”…, 32768) = -1 EPIPE (Broken pipe)  
— SIGPIPE {si\_signo=SIGPIPE, si\_code=SI\_USER, si\_pid=67738, si\_uid=1030} —  
rt\_sigprocmask(SIG\_BLOCK, [HUP INT PIPE TERM XCPU XFSZ], [HUP INT PIPE TERM XCPU XFSZ], 8) = 0  
rt\_sigprocmask(SIG\_SETMASK, [HUP INT PIPE TERM XCPU XFSZ], NULL, 8) = 0  
rt\_sigaction(SIGPIPE, {sa\_handler=SIG\_DFL, sa\_mask=[PIPE], sa\_flags=SA\_RESTORER|SA\_RESTART, sa\_restorer=0x7fb5a0713160}, {sa\_handler=0x55bc0c4beb2c, sa\_mask=[HUP INT PIPE TERM XCPU XFSZ], sa\_flags=SA\_RESTORER, sa\_restorer=0x7fb5a0713160}, 8) = 0  
rt\_sigprocmask(SIG\_BLOCK, ~[RTMIN RT\_1], [HUP INT PIPE TERM XCPU XFSZ], 8) = 0  
getpid() = 67738  
gettid() = 67738  
tgkill(67738, 67738, SIGPIPE) = 0  
rt\_sigprocmask(SIG\_SETMASK, [HUP INT PIPE TERM XCPU XFSZ], NULL, 8) = 0  
rt\_sigreturn({mask=[]}) = -1 EPIPE (Broken pipe)  
— SIGPIPE {si\_signo=SIGPIPE, si\_code=SI\_TKILL, si\_pid=67738, si\_uid=1030} —  
+++ killed by SIGPIPE +++[/QUOTE]

If I dont specify the number of threads, it runs like a charm.  
Also a different zip-tool like pbzip2 doesn’t have any problem at all.  
Prime95 runs without any problems for 24h, no bad sector detected, ram has been tested and is okay.

Also, on another server (same Software and OS - different hardware), there is no problem at all.

Anyone has a clue about this issue?

I also opend an issue for this problem at Github itself:

[https://github.com/madler/pigz/issues/68](https://github.com/madler/pigz/issues/68)

Thanks

---

<div class="post-metadata">

### Author: ![Mithras](https://avatars.discourse-cdn.com/v4/letter/m/f04885/32.png) [@Mithras](https://forums.suse.com/u/Mithras)
#### Post date: [April 2, 2019, 10:58am UTC](https://forums.suse.com/t/internal-threads-error-pigz/33748/2 "2019-04-02T10:58:58Z")

</div>

Wrong forum, its an SLES installation @mod could you please move the thread?